adding serde to the data model: triple

Thanks to @jonassmedegaard

Issue #842
Issue #843

.cargo/config.toml

@@ -1,137 +0,0 @@
-[build]
-rustflags = [
-    "-Wtrivial-casts",
-    "-Wtrivial-numeric-casts",
-    "-Wunsafe-code",
-    "-Wunused-lifetimes",
-    "-Wunused-qualifications",
-    # TODO: 1.63+ "-Wclippy::as-underscore",
-    # TODO: 1.65+ ""-Wclippy::bool-to-int-with-if",
-    "-Wclippy::borrow-as-ptr",
-    "-Wclippy::case-sensitive-file-extension-comparisons",
-    "-Wclippy::cast-lossless",
-    "-Wclippy::cast-possible-truncation",
-    "-Wclippy::cast-possible-wrap",
-    "-Wclippy::cast-precision-loss",
-    "-Wclippy::cast-ptr-alignment",
-    "-Wclippy::cast-sign-loss",
-    "-Wclippy::checked-conversions",
-    "-Wclippy::clone-on-ref-ptr",
-    "-Wclippy::cloned-instead-of-copied",
-    "-Wclippy::copy-iterator",
-    "-Wclippy::dbg-macro",
-    "-Wclippy::decimal-literal-representation",
-    "-Wclippy::default-trait-access",
-    "-Wclippy::default-union-representation",
-    # TODO: 1.61+ "-Wclippy::deref-by-slicing",
-    # TODO: 1.63+ "-Wclippy::doc-link-with-quotes",
-    # TODO: 1.62+ "-Wclippy::empty-drop",
-    "-Wclippy::empty-enum",
-    # TODO: on major version "-Wclippy::empty-structs-with-brackets",
-    "-Wclippy::enum-glob-use",
-    "-Wclippy::exit",
-    "-Wclippy::expect-used",
-    "-Wclippy::expl-impl-clone-on-copy",
-    "-Wclippy::explicit-deref-methods",
-    "-Wclippy::explicit-into-iter-loop",
-    "-Wclippy::explicit-iter-loop",
-    "-Wclippy::filter-map-next",
-    "-Wclippy::flat-map-option",
-    "-Wclippy::fn-to-numeric-cast-any",
-    # TODO: 1.62+ "-Wclippy::format-push-string",
-    "-Wclippy::from-iter-instead-of-collect",
-    "-Wclippy::get-unwrap",
-    "-Wclippy::if-not-else",
-    "-Wclippy::if-then-some-else-none",
-    "-Wclippy::implicit-clone",
-    "-Wclippy::inconsistent-struct-constructor",
-    "-Wclippy::index-refutable-slice",
-    "-Wclippy::inefficient-to-string",
-    "-Wclippy::inline-always",
-    "-Wclippy::inline-asm-x86-att-syntax",
-    "-Wclippy::inline-asm-x86-intel-syntax",
-    "-Wclippy::invalid-upcast-comparisons",
-    "-Wclippy::items-after-statements",
-    "-Wclippy::large-digit-groups",
-    # TODO: 1.68+ "-Wclippy::large-futures",
-    "-Wclippy::large-stack-arrays",
-    "-Wclippy::large-types-passed-by-value",
-    "-Wclippy::let-underscore-must-use",
-    "-Wclippy::let-unit-value",
-    "-Wclippy::linkedlist",
-    "-Wclippy::lossy-float-literal",
-    "-Wclippy::macro-use-imports",
-    "-Wclippy::manual-assert",
-    # TODO: 1.65+ "-Wclippy::manual-instant-elapsed",
-    # TODO: 1.67+ "-Wclippy::manual-let-else",
-    "-Wclippy::manual-ok-or",
-    # TODO: 1.65+ "-Wclippy::manual-string-new",
-    "-Wclippy::many-single-char-names",
-    "-Wclippy::map-unwrap-or",
-    "-Wclippy::match-bool",
-    "-Wclippy::match-same-arms",
-    "-Wclippy::match-wildcard-for-single-variants",
-    "-Wclippy::maybe-infinite-iter",
-    "-Wclippy::mem-forget",
-    # TODO: 1.63+ "-Wclippy::mismatching-type-param-order",
-    "-Wclippy::multiple-inherent-impl",
-    "-Wclippy::mut-mut",
-    "-Wclippy::mutex-atomic",
-    "-Wclippy::naive-bytecount",
-    "-Wclippy::needless-bitwise-bool",
-    "-Wclippy::needless-continue",
-    "-Wclippy::needless-pass-by-value",
-    "-Wclippy::no-effect-underscore-binding",
-    # TODO: 1.69+ "-Wclippy::no-mangle-with-rust-abi",
-    "-Wclippy::non-ascii-literal",
-    "-Wclippy::print-stderr",
-    "-Wclippy::print-stdout",
-    "-Wclippy::ptr-as-ptr",
-    "-Wclippy::range-minus-one",
-    "-Wclippy::range-plus-one",
-    "-Wclippy::rc-buffer",
-    "-Wclippy::rc-mutex",
-    "-Wclippy::redundant-closure-for-method-calls",
-    "-Wclippy::redundant-else",
-    "-Wclippy::redundant-feature-names",
-    "-Wclippy::ref-binding-to-reference",
-    "-Wclippy::ref-option-ref",
-    "-Wclippy::rest-pat-in-fully-bound-structs",
-    "-Wclippy::return-self-not-must-use",
-    "-Wclippy::same-functions-in-if-condition",
-    # TODO: strange failure on 1.60 "-Wclippy::same-name-method",
-    # TODO: 1.68+ "-Wclippy::semicolon-outside-block",
-    "-Wclippy::single-match-else",
-    "-Wclippy::stable-sort-primitive",
-    "-Wclippy::str-to-string",
-    "-Wclippy::string-add",
-    "-Wclippy::string-add-assign",
-    "-Wclippy::string-lit-as-bytes",
-    "-Wclippy::string-to-string",
-    # TODO: 1.67+ "-Wclippy::suspicious-xor-used-as-pow",
-    "-Wclippy::todo",
-    "-Wclippy::transmute-ptr-to-ptr",
-    "-Wclippy::trivially-copy-pass-by-ref",
-    "-Wclippy::try-err",
-    "-Wclippy::unicode-not-nfc",
-    "-Wclippy::unimplemented",
-    # TODO: 1.66+ "-Wclippy::uninlined-format-args",
-    # TODO: 1.70+ "-Wclippy::unnecessary-box-returns",
-    # TODO: 1.61+ "-Wclippy::unnecessary-join",
-    # TODO: 1.67+ "-Wclippy::unnecessary-safety-comment",
-    # TODO: 1.67+ "-Wclippy::unnecessary-safety-doc",
-    "-Wclippy::unnecessary-self-imports",
-    "-Wclippy::unnecessary-wraps",
-    "-Wclippy::unneeded-field-pattern",
-    "-Wclippy::unnested-or-patterns",
-    "-Wclippy::unreadable-literal",
-    "-Wclippy::unseparated-literal-suffix",
-    "-Wclippy::unused-async",
-    "-Wclippy::unused-self",
-    "-Wclippy::use-debug",
-    "-Wclippy::used-underscore-binding",
-    "-Wclippy::verbose-bit-mask",
-    "-Wclippy::verbose-file-reads",
-    "-Wclippy::wildcard-dependencies",
-    "-Wclippy::zero-sized-map-values",
-]

.clusterfuzzlite/Dockerfile

@@ -1,5 +1,4 @@
 FROM gcr.io/oss-fuzz-base/base-builder-rust:v1
-RUN apt-get update && apt-get install -y llvm-dev libclang-dev clang && apt-get clean && rm --recursive --force /var/lib/apt/lists/*
 COPY . $SRC/oxigraph
 WORKDIR oxigraph
 COPY .clusterfuzzlite/build.sh $SRC/

.clusterfuzzlite/build.sh

@@ -14,11 +14,17 @@ function build_seed_corpus() {
 
 
 cd "$SRC"/oxigraph
+git submodule init
+git submodule update
 cargo fuzz build -O --debug-assertions
-for TARGET in sparql_eval sparql_results_json sparql_results_tsv # sparql_results_xml https://github.com/tafia/quick-xml/issues/608
+for TARGET in sparql_eval sparql_results_json sparql_results_tsv sparql_results_xml n3 nquads trig rdf_xml
 do
   cp fuzz/target/x86_64-unknown-linux-gnu/release/$TARGET "$OUT"/
 done
 build_seed_corpus sparql_results_json srj
 build_seed_corpus sparql_results_tsv tsv
 build_seed_corpus sparql_results_xml srx
+build_seed_corpus n3 n3
+build_seed_corpus nquads nq
+build_seed_corpus trig trig
+build_seed_corpus rdf_xml rdf

.devcontainer/Dockerfile

@@ -1,21 +0,0 @@
-# See here for image contents: https://github.com/microsoft/vscode-dev-containers/tree/v0.241.1/containers/rust/.devcontainer/base.Dockerfile
-
-# [Choice] Debian OS version (use bullseye on local arm64/Apple Silicon): buster, bullseye
-ARG VARIANT="bullseye"
-FROM mcr.microsoft.com/vscode/devcontainers/rust:0-${VARIANT}
-
-# [Optional] Uncomment this section to install additional packages.
-RUN apt-get update && export DEBIAN_FRONTEND=noninteractive \
-    && apt-get -y install --no-install-recommends \
-        python3 \
-        python3-venv \
-        python-is-python3 \
-        libclang-dev
-
-ENV VIRTUAL_ENV=/opt/venv
-RUN python -m venv $VIRTUAL_ENV
-ENV PATH="$VIRTUAL_ENV/bin:$PATH"
-RUN pip install --no-cache-dir -r python/requirements.dev.txt
-
-# Change owner to the devcontainer user
-RUN chown -R 1000:1000 $VIRTUAL_ENV

.devcontainer/devcontainer.json

@@ -1,69 +0,0 @@
-// For format details, see https://aka.ms/devcontainer.json. For config options, see the README at:
-// https://github.com/microsoft/vscode-dev-containers/tree/v0.241.1/containers/rust
-{
-  "name": "Rust",
-  "build": {
-    "dockerfile": "Dockerfile",
-    "args": {
-      // Use the VARIANT arg to pick a Debian OS version: buster, bullseye
-      // Use bullseye when on local on arm64/Apple Silicon.
-      "VARIANT": "bullseye"
-    }
-  },
-  "runArgs": ["--cap-add=SYS_PTRACE", "--security-opt", "seccomp=unconfined"],
-
-  // Configure tool-specific properties.
-  "customizations": {
-    // Configure properties specific to VS Code.
-    "vscode": {
-      // Set *default* container specific settings.json values on container create.
-      "settings": {
-        "lldb.executable": "/usr/bin/lldb",
-        // VS Code don't watch files under ./target
-        "files.watcherExclude": {
-          "**/target/**": true
-        },
-        "rust-analyzer.checkOnSave.command": "clippy",
-
-        "python.defaultInterpreterPath": "/opt/venv/bin/python",
-        "python.linting.enabled": true,
-        "python.linting.pylintEnabled": true,
-        "python.formatting.autopep8Path": "/usr/local/py-utils/bin/autopep8",
-        "python.formatting.blackPath": "/usr/local/py-utils/bin/black",
-        "python.formatting.yapfPath": "/usr/local/py-utils/bin/yapf",
-        "python.linting.banditPath": "/usr/local/py-utils/bin/bandit",
-        "python.linting.flake8Path": "/usr/local/py-utils/bin/flake8",
-        "python.linting.mypyPath": "/usr/local/py-utils/bin/mypy",
-        "python.linting.pycodestylePath": "/usr/local/py-utils/bin/pycodestyle",
-        "python.linting.pydocstylePath": "/usr/local/py-utils/bin/pydocstyle",
-        "python.linting.pylintPath": "/opt/venv/bin/pylint",
-        "python.testing.pytestPath": "/opt/venv/bin/pytest"
-      },
-
-      // Add the IDs of extensions you want installed when the container is created.
-      "extensions": [
-        "vadimcn.vscode-lldb",
-        "mutantdino.resourcemonitor",
-        "rust-lang.rust-analyzer",
-        "tamasfe.even-better-toml",
-        "serayuzgur.crates",
-        "ms-python.python",
-        "ms-python.vscode-pylance",
-        "esbenp.prettier-vscode",
-        "stardog-union.stardog-rdf-grammars"
-      ]
-    }
-  },
-
-  // Use 'forwardPorts' to make a list of ports inside the container available locally.
-  // "forwardPorts": [],
-
-  // Use 'postCreateCommand' to run commands after the container is created.
-  "postCreateCommand": "git submodule update --init && cargo build",
-
-  // Comment out to connect as root instead. More info: https://aka.ms/vscode-remote/containers/non-root.
-  "remoteUser": "vscode",
-  "features": {
-    "python": "3.10"
-  }
-}

.github/ISSUE_TEMPLATE/bug_report.md

@@ -13,4 +13,4 @@ A clear and concise description of what the bug is.
 **To Reproduce**
 Steps to reproduce the behavior:
 1. Which version of Oxigraph are you using? On which platform?
-2. A command line or a code snippet that triggers the bug.
+2. A command-line or a code snippet that triggers the bug.

.github/actions/setup-rust/action.yml

@@ -0,0 +1,27 @@
+name: 'Setup Rust'
+description: 'Setup Rust using Rustup'
+inputs:
+  version:
+    description: 'Rust version to use. By default latest stable version'
+    required: false
+    default: 'stable'
+  component:
+    description: 'Rust extra component to install like clippy'
+    required: false
+  target:
+    description: 'Rust extra target to install like wasm32-unknown-unknown'
+    required: false
+runs:
+  using: "composite"
+  steps:
+    - run: rustup update
+      shell: bash
+    - run: rustup default ${{ inputs.version }}
+      shell: bash
+    - run: rustup component add ${{ inputs.component }}
+      shell: bash
+      if: ${{ inputs.component }}
+    - run: rustup target add ${{ inputs.target }}
+      shell: bash
+      if: ${{ inputs.target }}
+    - uses: Swatinem/rust-cache@v2

.github/workflows/artifacts.yml

@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main
+      - next
   release:
     types:
       - published
@@ -16,37 +17,39 @@ jobs:
   binary_linux:
     runs-on: ubuntu-20.04
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add aarch64-unknown-linux-gnu
+      - uses: ./.github/actions/setup-rust
+        with:
+          target: aarch64-unknown-linux-gnu
       - run: |
-          sudo apt update && sudo apt install -y g++-aarch64-linux-gnu
-          echo -e "\n\n[target.aarch64-unknown-linux-gnu]\nlinker = \"aarch64-linux-gnu-gcc\"" >> .cargo/config.toml
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo build --release
-        working-directory: ./server
-      - run: cargo build --release --target aarch64-unknown-linux-gnu
-        working-directory: ./server
+          sudo apt-get update && sudo apt-get install -y g++-aarch64-linux-gnu
+          mkdir .cargo
+          echo -e "[target.aarch64-unknown-linux-gnu]\nlinker = \"aarch64-linux-gnu-gcc\"" >> .cargo/config.toml
+      - run: cargo build --release --no-default-features --features rustls-native
+        working-directory: ./cli
+      - run: cargo build --release --target aarch64-unknown-linux-gnu --no-default-features --features rustls-native
+        working-directory: ./cli
         env:
           BINDGEN_EXTRA_CLANG_ARGS: --sysroot /usr/aarch64-linux-gnu
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: oxigraph_server_x86_64_linux_gnu
-          path: target/release/oxigraph_server
-      - uses: actions/upload-artifact@v3
+          name: oxigraph_x86_64_linux_gnu
+          path: target/release/oxigraph
+      - uses: actions/upload-artifact@v4
         with:
-          name: oxigraph_server_aarch64-linux_gnu
-          path: target/aarch64-unknown-linux-gnu/release/oxigraph_server
-      - run: mv target/release/oxigraph_server oxigraph_server_${{ github.event.release.tag_name }}_x86_64_linux_gnu
+          name: oxigraph_aarch64_linux_gnu
+          path: target/aarch64-unknown-linux-gnu/release/oxigraph
+      - run: mv target/release/oxigraph oxigraph_${{ github.event.release.tag_name }}_x86_64_linux_gnu
         if: github.event_name == 'release'
-      - run: mv target/aarch64-unknown-linux-gnu/release/oxigraph_server oxigraph_server_${{ github.event.release.tag_name }}_aarch64_linux_gnu
+      - run: mv target/aarch64-unknown-linux-gnu/release/oxigraph oxigraph_${{ github.event.release.tag_name }}_aarch64_linux_gnu
         if: github.event_name == 'release'
       - uses: softprops/action-gh-release@v1
         with:
           files: |
-            oxigraph_server_${{ github.event.release.tag_name }}_x86_64_linux_gnu
-            oxigraph_server_${{ github.event.release.tag_name }}_aarch64_linux_gnu
+            oxigraph_${{ github.event.release.tag_name }}_x86_64_linux_gnu
+            oxigraph_${{ github.event.release.tag_name }}_aarch64_linux_gnu
         if: github.event_name == 'release'
 
   binary_mac:
@@ -56,65 +59,64 @@ jobs:
       SDKROOT: '/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk'
       MACOSX_DEPLOYMENT_TARGET: '10.14'
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add aarch64-apple-darwin
-      - uses: Swatinem/rust-cache@v2
+      - uses: ./.github/actions/setup-rust
+        with:
+          target: aarch64-apple-darwin
       - run: cargo build --release
-        working-directory: ./server
+        working-directory: ./cli
       - run: cargo build --release --target aarch64-apple-darwin
-        working-directory: ./server
-      - uses: actions/upload-artifact@v3
+        working-directory: ./cli
+      - uses: actions/upload-artifact@v4
         with:
-          name: oxigraph_server_x86_64_apple
-          path: target/release/oxigraph_server
-      - uses: actions/upload-artifact@v3
+          name: oxigraph_x86_64_apple
+          path: target/release/oxigraph
+      - uses: actions/upload-artifact@v4
         with:
-          name: oxigraph_server_aarch64_apple
-          path: target/aarch64-apple-darwin/release/oxigraph_server
-      - run: mv target/release/oxigraph_server oxigraph_server_${{ github.event.release.tag_name }}_x86_64_apple
+          name: oxigraph_aarch64_apple
+          path: target/aarch64-apple-darwin/release/oxigraph
+      - run: mv target/release/oxigraph oxigraph_${{ github.event.release.tag_name }}_x86_64_apple
         if: github.event_name == 'release'
-      - run: mv target/aarch64-apple-darwin/release/oxigraph_server oxigraph_server_${{ github.event.release.tag_name }}_aarch64_apple
+      - run: mv target/aarch64-apple-darwin/release/oxigraph oxigraph_${{ github.event.release.tag_name }}_aarch64_apple
         if: github.event_name == 'release'
       - uses: softprops/action-gh-release@v1
         with:
           files: |
-            oxigraph_server_${{ github.event.release.tag_name }}_x86_64_apple
-            oxigraph_server_${{ github.event.release.tag_name }}_aarch64_apple
+            oxigraph_${{ github.event.release.tag_name }}_x86_64_apple
+            oxigraph_${{ github.event.release.tag_name }}_aarch64_apple
         if: github.event_name == 'release'
 
   binary_windows:
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
+      - uses: ./.github/actions/setup-rust
       - run: Remove-Item -LiteralPath "C:\msys64\" -Force -Recurse
       - run: cargo build --release
-        working-directory: ./server
-      - uses: actions/upload-artifact@v3
+        working-directory: ./cli
+      - uses: actions/upload-artifact@v4
         with:
-          name: oxigraph_server_x86_64_windows_msvc
-          path: target/release/oxigraph_server.exe
-      - run: mv target/release/oxigraph_server.exe oxigraph_server_${{ github.event.release.tag_name }}_x86_64_windows_msvc.exe
+          name: oxigraph_x86_64_windows_msvc
+          path: target/release/oxigraph.exe
+      - run: mv target/release/oxigraph.exe oxigraph_${{ github.event.release.tag_name }}_x86_64_windows_msvc.exe
         if: github.event_name == 'release'
       - uses: softprops/action-gh-release@v1
         with:
-          files: oxigraph_server_${{ github.event.release.tag_name }}_x86_64_windows_msvc.exe
+          files: oxigraph_${{ github.event.release.tag_name }}_x86_64_windows_msvc.exe
         if: github.event_name == 'release'
 
   python_sdist:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/setup-python@v5
         with:
           python-version: "3.12"
           cache: pip
@@ -123,18 +125,13 @@ jobs:
       - run: maturin build -m python/Cargo.toml
       - run: pip install --no-index --find-links=target/wheels/ pyoxigraph
       - run: rm -r target/wheels
-      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
+      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
         working-directory: ./python
       - run: maturin sdist -m python/Cargo.toml
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: pyoxigraph_source
+          name: pyoxigraph_sdist
           path: target/wheels/*.tar.gz
-      - run: pip install twine && twine upload target/wheels/*
-        env:
-          TWINE_USERNAME: __token__
-          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
-        if: github.event_name == 'release'
 
   wheel_linux:
     runs-on: ubuntu-latest
@@ -143,26 +140,21 @@ jobs:
         architecture: [ "x86_64", "aarch64" ]
     continue-on-error: true
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
       - uses: docker/setup-qemu-action@v2
         with:
           platforms: linux/${{ matrix.architecture }}
         if: github.event_name == 'release' && matrix.architecture != 'x86_64'
+      - uses: ./.github/actions/setup-rust
       - run: sed 's/%arch%/${{ matrix.architecture }}/g' .github/workflows/manylinux_build.sh | sed 's/%for_each_version%/${{ github.event_name == 'release' || '' }}/g' > .github/workflows/manylinux_build_script.sh
       - run: docker run -v "$(pwd)":/workdir --platform linux/${{ matrix.architecture }} quay.io/pypa/manylinux2014_${{ matrix.architecture }} /bin/bash /workdir/.github/workflows/manylinux_build_script.sh
         if: github.event_name == 'release' || matrix.architecture == 'x86_64'
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: pyoxigraph_wheel_x86_64_linux
+          name: pyoxigraph_${{ matrix.architecture }}_linux_gnu
           path: target/wheels/*.whl
-      - uses: pypa/gh-action-pypi-publish@release/v1
-        with:
-          user: __token__
-          password: ${{ secrets.PYPI_PASSWORD }}
-          packages-dir: target/wheels
-        if: github.event_name == 'release'
 
   wheel_linux_musl:
     runs-on: ubuntu-latest
@@ -171,26 +163,21 @@ jobs:
         architecture: [ "x86_64", "aarch64" ]
     continue-on-error: true
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
       - uses: docker/setup-qemu-action@v2
         with:
           platforms: linux/${{ matrix.architecture }}
         if: github.event_name == 'release' && matrix.architecture != 'x86_64'
+      - uses: ./.github/actions/setup-rust
       - run: sed 's/%arch%/${{ matrix.architecture }}/g' .github/workflows/musllinux_build.sh | sed 's/%for_each_version%/${{ github.event_name == 'release' || '' }}/g' > .github/workflows/musllinux_build_script.sh
-      - run: docker run -v "$(pwd)":/workdir --platform linux/${{ matrix.architecture }} quay.io/pypa/musllinux_1_1_${{ matrix.architecture }} /bin/bash /workdir/.github/workflows/musllinux_build_script.sh
+      - run: docker run -v "$(pwd)":/workdir --platform linux/${{ matrix.architecture }} quay.io/pypa/musllinux_1_2_${{ matrix.architecture }} /bin/bash /workdir/.github/workflows/musllinux_build_script.sh
         if: github.event_name == 'release' || matrix.architecture == 'x86_64'
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: pyoxigraph_wheel_x86_64_linux_musl
+          name: pyoxigraph_${{ matrix.architecture }}_linux_musl
           path: target/wheels/*.whl
-      - uses: pypa/gh-action-pypi-publish@release/v1
-        with:
-          user: __token__
-          password: ${{ secrets.PYPI_PASSWORD }}
-          packages-dir: target/wheels
-        if: github.event_name == 'release'
 
   wheel_mac:
     runs-on: macos-latest
@@ -199,78 +186,98 @@ jobs:
       SDKROOT: '/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk'
       MACOSX_DEPLOYMENT_TARGET: '10.14'
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add aarch64-apple-darwin
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - uses: ./.github/actions/setup-rust
+        with:
+          target: aarch64-apple-darwin
+      - uses: actions/setup-python@v5
         with:
-          python-version: "3.10"
+          python-version: "3.12"
           cache: pip
           cache-dependency-path: '**/requirements.dev.txt'
       - run: pip install -r python/requirements.dev.txt
-      - run: maturin build --release -m python/Cargo.toml --features abi3
+      - run: maturin build --release --features abi3
+        working-directory: ./python
       - run: pip install --no-index --find-links=target/wheels/ pyoxigraph
       - run: rm -r target/wheels
-      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
+      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
+        working-directory: ./python
+      - run: maturin build --release --target universal2-apple-darwin --features abi3
+        working-directory: ./python
+      - run: maturin build --release --features abi3
         working-directory: ./python
-      - run: maturin build --release --target universal2-apple-darwin -m python/Cargo.toml --features abi3
-      - run: maturin build --release -m python/Cargo.toml --features abi3
         if: github.event_name == 'release'
-      - run: maturin build --release --target aarch64-apple-darwin -m python/Cargo.toml --features abi3
+      - run: maturin build --release --target aarch64-apple-darwin --features abi3
+        working-directory: ./python
         if: github.event_name == 'release'
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: pyoxigraph_wheel_universal2_mac
+          name: pyoxigraph_macos
           path: target/wheels/*.whl
-      - run: pip install twine && twine upload target/wheels/*
-        env:
-          TWINE_USERNAME: __token__
-          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
-        if: github.event_name == 'release'
 
   wheel_windows:
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/setup-python@v5
         with:
-          python-version: "3.10"
+          python-version: "3.12"
           cache: pip
           cache-dependency-path: '**/requirements.dev.txt'
       - run: Remove-Item -LiteralPath "C:\msys64\" -Force -Recurse
       - run: pip install -r python/requirements.dev.txt
-      - run: maturin build --release -m python/Cargo.toml --features abi3
+      - run: maturin build --release --features abi3
+        working-directory: ./python
       - run: pip install --no-index --find-links=target/wheels/ pyoxigraph
       - run: rm -r target/wheels
-      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
+      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
         working-directory: ./python
       - run: maturin build --release -m python/Cargo.toml --features abi3
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
-          name: pyoxigraph_wheel_x86_64_windows
+          name: pyoxigraph_windows
           path: target/wheels/*.whl
-      - run: pip install twine && twine upload target/wheels/*
-        env:
-          TWINE_USERNAME: __token__
-          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
-        if: github.event_name == 'release'
+
+  publish_pypi:
+    if: github.event_name == 'release'
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/pyoxigraph
+    permissions:
+      id-token: write
+    needs:
+      - python_sdist
+      - wheel_windows
+      - wheel_mac
+      - wheel_linux
+      - wheel_linux_musl
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          pattern: pyoxigraph_*
+          path: dist
+          merge-multiple: true
+      - uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: dist
+          skip-existing: true
 
   npm_tarball:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo install wasm-pack || true
-      - uses: actions/setup-node@v3
+      - uses: ./.github/actions/setup-rust
+      - uses: taiki-e/install-action@v2
+        with: { tool: wasm-pack }
+      - uses: actions/setup-node@v4
         with:
           node-version: 16
           cache: npm
@@ -278,7 +285,7 @@ jobs:
           registry-url: https://registry.npmjs.org
       - run: npm run pack
         working-directory: ./js
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         with:
           name: oxigraph_wasm_npm
           path: js/*.tgz
@@ -291,27 +298,32 @@ jobs:
   docker:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - uses: docker/setup-buildx-action@v2
-      - uses: docker/login-action@v2
+      - uses: docker/setup-buildx-action@v3
+      - uses: docker/login-action@v3
         with:
           registry: ghcr.io
           username: ${{github.actor}}
           password: ${{secrets.GITHUB_TOKEN}}
-      - uses: docker/login-action@v2
+      - uses: docker/login-action@v3
         with:
           username: ${{ secrets.DOCKER_USERNAME }}
           password: ${{ secrets.DOCKER_PASSWORD }}
         if: github.event_name == 'release'
-      - uses: docker/metadata-action@v4
+      - uses: docker/metadata-action@v5
         id: docker_meta
         with:
           images: |
             ${{ github.repository }},enable=${{ github.event_name == 'release' }}
             ghcr.io/${{ github.repository }}
-      - uses: docker/build-push-action@v4
+          tags: |
+            type=ref,event=branch
+            type=ref,event=pr
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+      - uses: docker/build-push-action@v5
         with:
           context: .
           file: server/Dockerfile
@@ -327,7 +339,7 @@ jobs:
     if: github.event_name == 'release'
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
       - run: rustup update
@@ -343,27 +355,39 @@ jobs:
       - run: cargo publish
         working-directory: ./lib/oxrdf
         continue-on-error: true
+      - run: cargo publish
+        working-directory: ./lib/oxrdfxml
+        continue-on-error: true
+      - run: cargo publish
+        working-directory: ./lib/oxttl
+        continue-on-error: true
+      - run: cargo publish
+        working-directory: ./lib/oxrdfio
+        continue-on-error: true
       - run: cargo publish
         working-directory: ./lib/sparesults
         continue-on-error: true
       - run: cargo publish
         working-directory: ./lib/spargebra
         continue-on-error: true
+      - run: cargo publish
+        working-directory: ./lib/sparopt
+        continue-on-error: true
       - run: cargo publish
         working-directory: ./lib/sparql-smith
         continue-on-error: true
       - run: cargo publish
-        working-directory: ./lib
+        working-directory: ./lib/oxigraph
         continue-on-error: true
       - run: cargo publish
-        working-directory: ./server
+        working-directory: ./cli
 
   homebrew:
     if: "github.event_name == 'release' && !contains('-', github.event.release.tag_name)"
     runs-on: ubuntu-latest
     needs: full_archive
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           repository: oxigraph/homebrew-oxigraph
           token: ${{ secrets.FULL_ACCESS_TOKEN }}
@@ -383,7 +407,7 @@ jobs:
     if: github.event_name == 'release'
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
       - run: |

.github/workflows/install_rocksdb.sh

@@ -0,0 +1,11 @@
+if [ -f "rocksdb" ]
+then
+  cd rocksdb || exit
+else
+  git clone https://github.com/facebook/rocksdb.git
+  cd rocksdb || exit
+  git checkout v8.0.0
+  make shared_lib
+fi
+sudo make install-shared
+sudo ldconfig /usr/local/lib

.github/workflows/manylinux_build.sh

@@ -8,17 +8,17 @@ chmod +x rustup-init
 source "$HOME/.cargo/env"
 export PATH="${PATH}:/opt/python/cp37-cp37m/bin:/opt/python/cp38-cp38/bin:/opt/python/cp39-cp39/bin:/opt/python/cp310-cp310/bin:/opt/python/cp311-cp311/bin"
 cd python
-python3.10 -m venv venv
+python3.12 -m venv venv
 source venv/bin/activate
 pip install -r requirements.dev.txt
-maturin develop --release -m Cargo.toml
-python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
-maturin build --release -m Cargo.toml --features abi3 --compatibility manylinux2014
+maturin develop --release
+python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
+maturin build --release --features abi3 --compatibility manylinux2014
 if [ %for_each_version% ]; then
-  for VERSION in 7 8 9 10 11 12; do
-    maturin build --release -m Cargo.toml --interpreter "python3.$VERSION" --compatibility manylinux2014
+  for VERSION in 8 9 10 11 12; do
+    maturin build --release --interpreter "python3.$VERSION" --compatibility manylinux2014
   done
   for VERSION in 9 10; do
-    maturin build --release -m Cargo.toml --interpreter "pypy3.$VERSION" --compatibility manylinux2014
+    maturin build --release --interpreter "pypy3.$VERSION" --compatibility manylinux2014
   done
 fi

.github/workflows/musllinux_build.sh

@@ -6,14 +6,14 @@ chmod +x rustup-init
 source "$HOME/.cargo/env"
 export PATH="${PATH}:/opt/python/cp37-cp37m/bin:/opt/python/cp38-cp38/bin:/opt/python/cp39-cp39/bin:/opt/python/cp310-cp310/bin:/opt/python/cp311-cp311/bin"
 cd python
-python3.10 -m venv venv
+python3.12 -m venv venv
 source venv/bin/activate
 pip install -r requirements.dev.txt
-maturin develop --release -m Cargo.toml
-python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
-maturin build --release -m Cargo.toml --features abi3 --compatibility musllinux_1_1
+maturin develop --release
+python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
+maturin build --release --features abi3 --compatibility musllinux_1_2
 if [ %for_each_version% ]; then
-  for VERSION in 7 8 9 10 11 12; do
-    maturin build --release -m Cargo.toml --interpreter "python3.$VERSION" --compatibility musllinux_1_1
+  for VERSION in 8 9 10 11 12; do
+    maturin build --release --interpreter "python3.$VERSION" --compatibility musllinux_1_2
   done
 fi

.github/workflows/tests.yml

@@ -4,6 +4,7 @@ on:
   pull_request:
     branches:
       - main
+      - next
   schedule:
     - cron: "12 3 * * *"
 
@@ -15,209 +16,259 @@ jobs:
   fmt:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: rustup update && rustup component add rustfmt
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/setup-rust
+        with:
+          component: rustfmt
       - run: cargo fmt -- --check
 
   clippy:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: 1.76.0
+          component: clippy
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
         working-directory: ./lib/oxsdatatypes
-      - run: cargo clippy
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
         working-directory: ./lib/oxrdf
-      - run: cargo clippy
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./lib/oxrdfxml
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./lib/oxttl
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./lib/oxrdfio
+      - run: cargo clippy --all-targets --features async-tokio -- -D warnings -D clippy::all
+        working-directory: ./lib/oxrdfio
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
         working-directory: ./lib/sparesults
-      - run: cargo clippy
+      - run: cargo clippy --all-targets --features async-tokio -- -D warnings -D clippy::all
+        working-directory: ./lib/sparesults
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
         working-directory: ./lib/spargebra
-      - run: cargo clippy --all-targets --all-features
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./lib/sparopt
+      - run: cargo clippy --all-targets --no-default-features -- -D warnings -D clippy::all
+        working-directory: ./lib/oxigraph
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./lib/oxigraph
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./python
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./cli
+      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
+        working-directory: ./testsuite
 
   clippy_wasm_js:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add wasm32-unknown-unknown && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy --lib --tests --target wasm32-unknown-unknown
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: 1.76.0
+          target: wasm32-unknown-unknown
+          component: clippy
+      - run: cargo clippy --lib --tests --target wasm32-unknown-unknown -- -D warnings -D clippy::all
         working-directory: ./js
 
   clippy_wasi:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add wasm32-wasi && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy --lib --tests --target wasm32-wasi
-        working-directory: ./lib
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: 1.76.0
+          target: wasm32-wasi
+          component: clippy
+      - run: cargo clippy --lib --tests --target wasm32-wasi -- -D warnings -D clippy::all
+        working-directory: ./lib/oxigraph
+      - run: cargo clippy --target wasm32-wasi --features abi3 --no-default-features -- -D warnings -D clippy::all
+        working-directory: ./python
 
-  clippy_msv:
+  clippy_wasm_emscripten:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup override set 1.60.0 && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy --  -D warnings -D clippy::all
-        working-directory: ./lib/oxsdatatypes
-      - run: cargo clippy --  -D warnings -D clippy::all
-        working-directory: ./lib/oxrdf
-      - run: cargo clippy -- -D warnings -D clippy::all
-        working-directory: ./lib/sparesults
-      - run: cargo clippy -- -D warnings -D clippy::all
-        working-directory: ./lib/spargebra
-      - run: cargo clippy --all-targets -- -D warnings -D clippy::all
-        working-directory: ./server
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: 1.76.0
+          target: wasm32-unknown-emscripten
+          component: clippy
+      - run: cargo clippy --lib --tests --target wasm32-unknown-emscripten -- -D warnings -D clippy::all
+        working-directory: ./lib/oxigraph
+      - run: cargo clippy --target wasm32-unknown-emscripten --features abi3 -- -D warnings -D clippy::all
+        working-directory: ./python
 
-  clippy_msv_wasm_js:
+  clippy_wasm_unknown:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup override set 1.60.0 && rustup target add wasm32-unknown-unknown && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy --lib --tests --target wasm32-unknown-unknown -- -D warnings -D clippy::all
-        working-directory: ./js
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: 1.76.0
+          target: wasm32-unknown-unknown
+          component: clippy
+      - run: cargo clippy --lib --tests --target wasm32-unknown-unknown --features getrandom/custom --features oxsdatatypes/custom-now -- -D warnings -D clippy::all
+        working-directory: ./lib/oxigraph
+
+  deny:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: taiki-e/install-action@v2
+        with: { tool: cargo-deny }
+      - run: cargo deny check
 
-  clippy_msv_wasi:
+  semver_checks:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup override set 1.60.0 && rustup target add wasm32-wasi && rustup component add clippy
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo clippy --lib --tests --target wasm32-wasi -- -D warnings -D clippy::all
-        working-directory: ./lib
+      - uses: ./.github/actions/setup-rust
+      - uses: taiki-e/install-action@v2
+        with: { tool: cargo-semver-checks }
+      - uses: actions/cache@v4
+        with:
+          path: rocksdb
+          key: ${{ runner.os }}-rocksdb-8.0.0
+      - run: bash .github/workflows/install_rocksdb.sh
+      - run: cargo semver-checks check-release --exclude oxrocksdb-sys --exclude oxigraph-js --exclude pyoxigraph --exclude oxigraph-testsuite --exclude oxigraph-cli
 
-  deny:
+  test_linux_x86_64:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo install cargo-deny || true
-      - run: cargo deny check
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: ./.github/actions/setup-rust
+      - run: cargo test
 
-  semver_checks:
+  test_linux_i686:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo install cargo-semver-checks || true
-      - run: cargo semver-checks check-release --exclude oxrocksdb-sys --exclude oxigraph_js --exclude pyoxigraph --exclude oxigraph_testsuite --exclude oxigraph_server
+      - uses: ./.github/actions/setup-rust
+        with:
+          target: i686-unknown-linux-gnu
+      - run: sudo apt-get update && sudo apt-get install -y g++-multilib
+      - run: cargo test --target i686-unknown-linux-gnu --no-default-features --features http-client-rustls-native
+        working-directory: ./lib/oxigraph
 
-  test_linux:
+  test_linux_msv:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo test --all-features
-        env:
-          RUST_BACKTRACE: 1
+      - name: Read MSRV from Cargo.toml
+        id: metadata
+        run: echo "rust-version=$(sed -ne 's/rust-version *= *\"\(.*\)\"/\1/p' Cargo.toml)" >> $GITHUB_OUTPUT
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: ${{ steps.metadata.outputs.rust-version }}
+      - run: rustup toolchain install nightly
+      - run: rm Cargo.lock && cargo +nightly update -Z direct-minimal-versions && cargo update -p bumpalo --precise 3.14.0
+      - run: cargo test
 
   test_linux_latest:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
+      - uses: ./.github/actions/setup-rust
       - run: rm Cargo.lock && cargo update
       - run: cargo test
 
-  address_sanitizer:
+  test_linux_address_sanitizer:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup override set nightly
-      - run: sudo apt install -y llvm
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo test --tests --target x86_64-unknown-linux-gnu --workspace --exclude pyoxigraph --exclude oxigraph_testsuite --exclude oxigraph_server
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: nightly
+      - run: sudo apt-get update && sudo apt-get install -y llvm
+      - run: cargo test --tests --target x86_64-unknown-linux-gnu --workspace --exclude pyoxigraph --exclude oxigraph-testsuite --exclude oxigraph-cli
         env:
-          RUST_BACKTRACE: 1
           RUSTFLAGS: -Z sanitizer=address
 
+  test_linux_dynamic_linking:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/cache@v4
+        with:
+          path: rocksdb
+          key: ${{ runner.os }}-rocksdb-8.0.0
+      - run: bash .github/workflows/install_rocksdb.sh
+      - run: cargo test --tests --features oxrocksdb-sys/pkg-config
+
   test_windows:
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
+      - uses: ./.github/actions/setup-rust
       - run: Remove-Item -LiteralPath "C:\msys64\" -Force -Recurse
-      - run: cargo test --all-features
-        env:
-          RUST_BACKTRACE: 1
+      - run: cargo test
 
   test_wasi:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update && rustup target add wasm32-wasi
-      - uses: Swatinem/rust-cache@v2
-      - uses: taiki-e/install-action@wasmtime
-      - run: cargo install cargo-wasi || true
-      - run: cargo wasi test --workspace --exclude oxigraph_js --exclude oxigraph_server --exclude oxigraph_testsuite --exclude oxrocksdb-sys --exclude pyoxigraph
-        env:
-          RUST_BACKTRACE: 1
+      - uses: ./.github/actions/setup-rust
+        with:
+          target: wasm32-wasi
+      - uses: taiki-e/install-action@v2
+        with: { tool: "wasmtime,cargo-wasi" }
+      - run: cargo wasi test --workspace --exclude oxigraph-js --exclude oxigraph-cli --exclude oxigraph-testsuite --exclude oxrocksdb-sys --exclude pyoxigraph
 
   rustdoc:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo doc --all-features
-        working-directory: ./lib
-
-  rustdoc_msrv:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v3
+      - uses: ./.github/actions/setup-rust
         with:
-          submodules: true
-      - run: rustup update && rustup override set 1.60.0
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo doc --all-features
-        working-directory: ./lib
+          version: 1.76.0
+      - run: cargo doc
         env:
           RUSTDOCFLAGS: -D warnings
 
   js:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo install wasm-pack || true
-      - uses: actions/setup-node@v3
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/setup-rust
+      - uses: taiki-e/install-action@v2
+        with: { tool: wasm-pack }
+      - uses: actions/setup-node@v4
         with:
-          node-version: 16
+          node-version: 18
           cache: npm
           cache-dependency-path: "js/package.json"
       - run: npm install
@@ -228,48 +279,72 @@ jobs:
   python:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/setup-python@v5
         with:
-          python-version: "3.10"
+          python-version: "3.12"
           cache: pip
           cache-dependency-path: '**/requirements.dev.txt'
       - run: pip install -r python/requirements.dev.txt
-      - run: python -m black --check --diff --color .
-        working-directory: ./python
       - run: maturin build -m python/Cargo.toml
       - run: pip install --no-index --find-links=target/wheels/ pyoxigraph
       - run: rm -r target/wheels
       - run: python -m unittest
         working-directory: ./python/tests
-      - run: sphinx-build -M doctest . build
-        working-directory: ./python/docs
       - run: sphinx-build -M html . build
         working-directory: ./python/docs
-      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --black
+      - run: python generate_stubs.py pyoxigraph pyoxigraph.pyi --ruff
         working-directory: ./python
       - run: python -m mypy.stubtest pyoxigraph  --allowlist=mypy_allowlist.txt
         working-directory: ./python
       - run: python -m mypy generate_stubs.py tests --strict
         working-directory: ./python
-      - run: python -m ruff check .
+      - run: python -m ruff format --check .
+        working-directory: ./python
+      - run: python -m ruff check --output-format=github .
+        working-directory: ./python
+      - run: sphinx-lint docs
         working-directory: ./python
 
   python_msv:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - name: Read MSRV from Cargo.toml
+        id: metadata
+        run: echo "rust-version=$(sed -ne 's/rust-version *= *\"\(.*\)\"/\1/p' Cargo.toml)" >> $GITHUB_OUTPUT
+      - uses: ./.github/actions/setup-rust
         with:
-          python-version: "3.7"
+          version: ${{ steps.metadata.outputs.rust-version }}
+      - run: rustup toolchain install nightly
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.8"
+          cache: pip
+          cache-dependency-path: '**/requirements.dev.txt'
+      - run: pip install -r python/requirements.dev.txt
+      - run: rm Cargo.lock && cargo +nightly update -Z direct-minimal-versions
+      - run: maturin build -m python/Cargo.toml
+      - run: pip install --no-index --find-links=target/wheels/ pyoxigraph
+      - run: rm -r target/wheels
+      - run: python -m unittest
+        working-directory: ./python/tests
+
+  python_pypy:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "pypy3.10"
           cache: pip
           cache-dependency-path: '**/requirements.dev.txt'
       - run: pip install -r python/requirements.dev.txt
@@ -282,14 +357,13 @@ jobs:
   python_windows:
     runs-on: windows-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           submodules: true
-      - run: rustup update
-      - uses: Swatinem/rust-cache@v2
-      - uses: actions/setup-python@v4
+      - uses: ./.github/actions/setup-rust
+      - uses: actions/setup-python@v5
         with:
-          python-version: "3.10"
+          python-version: "3.12"
           cache: pip
           cache-dependency-path: '**/requirements.dev.txt'
       - run: pip install "maturin~=1.0"
@@ -299,28 +373,55 @@ jobs:
       - run: python -m unittest
         working-directory: ./python/tests
 
+  python_pyodide:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: '**/requirements.dev.txt'
+      - uses: ./.github/actions/setup-rust
+        with:
+          version: nightly
+          target: wasm32-unknown-emscripten
+      - run: |
+          pip install pyodide-build
+          echo EMSCRIPTEN_VERSION=$(pyodide config get emscripten_version) >> $GITHUB_ENV
+      - uses: mymindstorm/setup-emsdk@v13
+        with:
+          version: ${{ env.EMSCRIPTEN_VERSION }}
+      - run: pyodide build
+        working-directory: ./python
+      - run: |
+          pyodide venv venv
+          source venv/bin/activate
+          pip install --no-index --find-links=../dist/ pyoxigraph
+          python -m unittest
+        working-directory: ./python/tests
+
   typos:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - uses: Swatinem/rust-cache@v2
-      - run: cargo install typos-cli || true
+      - uses: actions/checkout@v4
+      - uses: taiki-e/install-action@v2
+        with: { tool: typos-cli }
       - run: typos
 
   clang_fmt:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: sudo apt install -y clang-format
+      - uses: actions/checkout@v4
+      - run: sudo apt-get update && sudo apt-get install -y clang-format
       - run: clang-format --Werror --dry-run oxrocksdb-sys/api/*
 
   fuzz_changes:
     if: github.event_name == 'pull_request'
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-        with:
-          submodules: true
       - uses: google/clusterfuzzlite/actions/build_fuzzers@v1
         with:
           language: rust
@@ -330,7 +431,7 @@ jobs:
       - uses: google/clusterfuzzlite/actions/run_fuzzers@v1
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
-          fuzz-seconds: 600
+          fuzz-seconds: 900
           mode: code-change
           sanitizer: address
           minimize-crashes: true
@@ -341,9 +442,6 @@ jobs:
     if: github.event_name != 'pull_request'
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-        with:
-          submodules: true
       - uses: google/clusterfuzzlite/actions/build_fuzzers@v1
         with:
           language: rust
@@ -352,7 +450,7 @@ jobs:
       - uses: google/clusterfuzzlite/actions/run_fuzzers@v1
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
-          fuzz-seconds: 3600
+          fuzz-seconds: 7200
           mode: batch
           sanitizer: address
           minimize-crashes: true
@@ -387,12 +485,52 @@ jobs:
   shellcheck:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: sudo apt install -y shellcheck
+      - uses: actions/checkout@v4
+      - run: sudo apt-get update && sudo apt-get install -y shellcheck
       - run: git grep -l '^#\( *shellcheck \|!\(/bin/\|/usr/bin/env \)\(sh\|bash\|dash\|ksh\)\)' | xargs shellcheck
 
   spec_links:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - run: python test_spec_links.py
+      - uses: actions/checkout@v4
+      - run: python lints/test_spec_links.py
+
+  debian_compatibility:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/setup-rust
+      - run: python lints/test_debian_compatibility.py
+
+  codspeed:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: ./.github/actions/setup-rust
+      - uses: taiki-e/install-action@v2
+        with: { tool: cargo-codspeed }
+      - run: cargo codspeed build -p oxigraph --features http-client-native-tls
+      - run: cargo codspeed build -p oxigraph-testsuite
+      - uses: CodSpeedHQ/action@v2
+        with:
+          run: cargo codspeed run
+          token: ${{ secrets.CODSPEED_TOKEN }}
+
+  codecov:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - uses: ./.github/actions/setup-rust
+      - uses: taiki-e/install-action@v2
+        with: { tool: cargo-llvm-cov }
+      - run: cargo llvm-cov --codecov --output-path codecov.json
+      - uses: codecov/codecov-action@v4
+        with:
+          files: codecov.json
+          flags: rust
+          fail_ci_if_error: true
+          token: ${{ secrets.CODECOV_TOKEN }}

.gitmodules

@@ -7,9 +7,13 @@
 [submodule "bench/bsbm-tools"]
 	path = bench/bsbm-tools
 	url = https://github.com/Tpt/bsbm-tools.git
-[submodule "oxrocksdb-sys/rocksdb"]
-	path = oxrocksdb-sys/rocksdb
-	url = https://github.com/oxigraph/rocksdb.git
 [submodule "oxrocksdb-sys/lz4"]
 	path = oxrocksdb-sys/lz4
 	url = https://github.com/lz4/lz4.git
+[submodule "testsuite/N3"]
+	path = testsuite/N3
+	url = https://github.com/w3c/N3.git
+	branch = master
+[submodule "testsuite/rdf-canon"]
+	path = testsuite/rdf-canon
+	url = https://github.com/w3c/rdf-canon.git

.readthedocs.yaml

@@ -7,7 +7,7 @@ build:
   os: "ubuntu-22.04"
   tools:
     python: "3"
-    rust: "1.64"
+    rust: "1.70"
   apt_packages:
     - clang
 

CHANGELOG.md

@@ -1,3 +1,123 @@
+## [0.4.0-alpha.6] - 2024-03-25
+
+### Changed
+- Fixes compatibility with OxIRI 0.2.3.
+
+
+## [0.4.0-alpha.5] - 2024-03-23
+
+### Added
+- Python: `Dataset` class
+
+### Changed
+- Rust: `Dataset::canonicalize` and `Graph::canonicalize` takes for input a `CanonicalizationAlgorithm` to set which algorithm to use.
+- Upgrades RocksDB to 9.0.0
+- JS: Drops NodeJS 12-16 and older web browsers (Edge before Chromium...) support.
+
+
+## [0.4.0-alpha.4] - 2024-03-07
+
+### Added
+- Rust: `From<spargebra::Update` on `Update`.
+- `sparesults`: Tokio Async readers for SPARQL query results (XML, JSON and TSV).
+- `oxrdf`: `Term::from_terms` constructor.
+- JS: options to set query and update base IRI and query union default graph.
+
+## Changed
+- Uses RocksDB atomic multi-columns flush.
+- Rust: RocksDB is optional but enabled by default.
+- Upgrades RocksDB to 8.11.3.
+- Rust: `oxigraph` crate has been moved to `lib/oxigraph`.
+- Rust: `QueryResults::write` returns the `Write` impl and not `()`.
+- Rust: use `thierror` for errors.
+- Rust: rename `oxrdfio::{Parse,Syntax}Error` to `oxrdfio::Rdf{Parse,Syntax}Error`,
+  `oxrdfxml::{Parse,Syntax}Error` to `oxrdfxml::RdfXml{Parse,Syntax}Error`,
+  `oxttl::{Parse,Syntax}Error` to `oxttl::Turtle{Parse,Syntax}Error`,
+  `sparesults::{Parse,Syntax}Error` to `sparesults::QueryResults{Parse,Syntax}Error` and
+  `spargebra::SyntaxError` to `spargebra::SparqlSyntaxError`.
+
+
+## [0.4.0-alpha.3] - 2024-01-25
+
+### Added
+- `oxttl`: expose base IRIs.
+- `oxttl`: allows to inject prefixes for serialization.
+- `oxrdf`: `vocab::geosparql::WKT_LITERAL`.
+
+### Changed
+- Turtle: Fixes parsing bug with escaped dot at the end of local name.
+- `oxttl`: Changes `prefixes` getter return type. 
+- JS: simplify build.
+- Python: uses rustls by default all platforms that are not Windows/macOS/iOS/WASM.
+- Strips debug info of the Rust std library in release build.
+
+
+## [0.4.0-alpha.2] - 2024-01-08
+
+### Added
+- i686 linux support
+
+### Changed
+- Docker: fixes Docker image Glib version error.
+- Docker: tags now use semver e.g. `0.3.22` and not `v0.3.22`. Preversions are also not tagged `latest` anymore.
+- Python: `QuerySolution` is now thread safe.
+
+
+## [0.4.0-alpha.1] - 2024-01-03
+
+### Added
+- `sparopt` crate: A new still quite naive query optimizer.
+- `oxttl` crate: A N-Triples/N-Quads/Turtle/TriG/N3 parser and serializer compatible with Tokio.
+- `oxrdfxml` crate: A RDF/XML parser and serializer compatible with Tokio.
+- `oxrdfio` crate: A stand-alone crate with oxigraph I/O related APIs.
+- Rust: SPARQL results I/O is now exposed in the `oxigraph` crate (`oxigraph::sparql::results` module).
+- Rust: It is now possible to dynamically link rocksdb with the `rocksdb-pkg-config` feature.
+- Python: error location is now included in some `SyntaxError` exceptions.
+- Python: the file type can be guessed from the file path extension during parsing and serialization.
+- Python: the serialization method returns a `bytes` value if no output-related argument is given.
+- Python: SPARQL query results I/O is now exposed (`parse_query_results` function and `.serialize` method).
+- Python: `RdfFormat` and `QueryResultsFormat` enum to encode supported formats.
+- CLI: a `convert` command to convert RDF file between different formats.
+
+### Removed
+- Rust: automated flush at the end of serialization. This should be done explicitly now.
+- oxsdatatypes: Deprecated methods.
+- Python: 3.7 and Musl linux 1.1 support.
+- Python: `GraphName.value`.
+
+### Changed
+- SPARQL: a digit after `.` is now required for `xsd:decimal`.
+- SPARQL: calendar subtraction returns `xsd:dayTimeDuration` and not `xsd:duration`.
+- SPARQL: Unicode escapes (`\u` and `\U`) are now only supported in IRIs and strings and not everywhere.
+- Literal serialization now produces canonical N-Triples according to the RDF 1.2 and RDF Dataset Canonicalization drafts
+- Rust: MSRV is now 1.70.
+- Rust Makes `GraphName` implement `Default`.
+- Rust: `wasm32-unknown-unknown` does not assumes JS platform by default. Enable the `js` feature for that.
+- Rust: Parsers take `Read` and not `BufRead` for input.
+- Rust: `GraphFormat` and `DatasetFormat` have been merged into `RdfFormat`.
+- Rust: `GraphParser` and `DatasetParser` have been merged into `RdfParser`.
+- Rust: `GraphSerializer` and `DatasetSerializer` have been merged into `RdfSerializer`.
+- Rust: query results are now `Send` and `Sync`.
+- Rust: `Store.load_graph` and `Store.load_dataset` have been merged into a `load_from_read` method.
+- Rust: `Store.dump_graph` and `Store.dump_dataset` have been renamed to `dump_graph_to_write` and `dump_to_write`.
+- Rust: `BulkLoader.set_*` methods have been renamed to `BulkLoader.with_*`.
+- oxsdatatypes: pass by-values instead of by-reference parameters when relevant.
+- oxsdatatypes: error types have been redesigned.
+- oxsdatatypes: return an error when building not serializable duration (year-month and day-time of opposite signs).
+- sparesults: renames some methods to move closer to the new oxrdfio crate.
+- Python: raise `OSError` instead of `IOError` on OS errors.
+- Python: the `mime_type` parameter have been renamed to `format`.
+- Python: boolean SPARQL results are now encoded with a `QueryBoolean` class and not a simple `bool`.
+- Python: a `path` parameter has been added to all I/O method to read from a file.
+  The existing `input` parameter now consider `str` values to be a serialization to parse.
+- JS: the `mime_type` parameter have been renamed to `format`.
+- CLI: the `oxigraph_server` binary has been renamed to `oxigraph`.
+- CLI: the `--location` argument is now part of sub-commands where it is relevant.
+  `oxigraph_server --location foo serve` is not possible anymore.
+  One need to write `oxigraph serve --location foo`.
+- CLI: is is now possible to upload gzip encoded files to the HTTP API with the `Content-Encoding: gzip` header.
+
+
 ## [0.3.22] - 2023-11-29
 
 ### Changed
@@ -321,7 +441,7 @@
 
 ### Added
 - [RDF-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html) support. `Triple` is now a possible `Term`. Serialization formats and SPARQL support have been updated to match the [latest version of the specification draft](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html).
-- Fast data bulk load with the `Store` `bulk_load_dataset` and `bulk_load_graph` methods and a special command line option of the server.
+- Fast data bulk load with the `Store` `bulk_load_dataset` and `bulk_load_graph` methods and a special command-line option of the server.
 - It is now possible to quickly backup the database using the `backup` method.
 - Rust: `*Syntax::from_extension` to easy guess a graph/dataset/sparql result format from a file extension.
 - Rust: Custom SPARQL functions are now supported using `QueryOptions::with_custom_function`.
@@ -337,7 +457,7 @@
 - SPARQL: The parser now validates more carefully the inputs following the SPARQL specification and test suite.
 - SPARQL: Variable scoping was buggy with "FILTER EXISTS". It is now fixed.
 - Rust: RDF model, SPARQL parser and SPARQL result parsers have been moved to stand-alone reusable libraries.
-- Rust: HTTPS is not supported by default with the `http_client` option. You need to enable the `native-tls` or the `rustls` feature of the `oxhttp` crate to enable a TSL layer.
+- Rust: HTTPS is not supported by default with the `http_client` option. You need to enable the `native-tls` or the `rustls` feature of the `oxhttp` crate to enable a TLS layer.
 - Rust: The error types have been cleaned.
   Most of the `Store` methods now return a `StorageError` that is more descriptive than the previous `std::io::Error`.
   The new error type all implements `Into<std::io::Error>` for easy conversion.
@@ -345,7 +465,7 @@
   It is The used type of the `subject` field of the `Triple` and `Quad` structs.
 - Rust: The SPARQL algebra is not anymore publicly exposed in the `oxigraph` crate. The new `oxalgebra` crate exposes it.
 - Rust: `UpdateOptions` API have been rewritten. It can now be built using `From<QueryOptions>` or `Default`.
-- Server: The command line API has been redesign. See the [server README](server/README.md) for more information.
+- Server: The command-line API has been redesign. See the [server README](server/README.md) for more information.
 - Server: The HTTP implementation is now provided by [`oxhttp`](https://github.com/oxigraph/oxhttp).
 - Server: The HTTP response bodies are now generated on the fly instead of being buffered.
 - Python: The `SledStore` and `MemoryStore` classes have been removed in favor of the `Store` class.

Cargo.toml

@@ -1,22 +1,272 @@
 [workspace]
 members = [
     "js",
-    "lib",
+    "lib/oxigraph",
     "lib/oxrdf",
+    "lib/oxrdfio",
+    "lib/oxrdfxml",
     "lib/oxsdatatypes",
-    "lib/spargebra",
+    "lib/oxttl",
     "lib/sparesults",
+    "lib/spargebra",
+    "lib/sparopt",
     "lib/sparql-smith",
-    "oxrocksdb-sys",
-    "python",
-    "server",
     "testsuite"
 ]
 resolver = "2"
 
+[workspace.package]
+version = "0.4.0-alpha.7-dev"
+authors = ["Tpt <thomas@pellissier-tanon.fr>"]
+license = "MIT OR Apache-2.0"
+edition = "2021"
+rust-version = "1.70"
+
+[workspace.dependencies]
+rocksdb = {git = "https://git.nextgraph.org/NextGraph/rust-rocksdb.git", branch = "master", features = [  ] }
+serde = { version = "1.0.142", features = ["derive"] }
+anyhow = "1.0.72"
+arbitrary = "1.3"
+assert_cmd = "2.0"
+assert_fs = "1.0"
+bindgen = ">=0.60, <0.70"
+cc = "1.0.73"
+clap = "4.0"
+codspeed-criterion-compat = "2.3.3"
+console_error_panic_hook = "0.1.7"
+digest = "0.10"
+flate2 = "1.0"
+getrandom = "0.2.8"
+hex = "0.4"
+js-sys = "0.3.60"
+json-event-parser = "0.2.0-alpha.2"
+md-5 = "0.10"
+memchr = "2.5"
+oxilangtag = "0.1"
+oxiri = "0.2.3"
+peg = "0.8"
+pkg-config = "0.3.25"
+predicates = ">=2.0, <4.0"
+pyo3 = "0.21.0"
+quick-xml = ">=0.29, <0.32"
+rand = "0.8"
+rayon-core = "1.11"
+regex = "1.7"
+sha1 = "0.10"
+sha2 = "0.10"
+siphasher = ">=0.3, <2.0"
+text-diff = "0.4"
+thiserror = "1.0.50"
+time = "0.3"
+tokio = "1.29"
+url = "2.4"
+wasm-bindgen = "0.2.83"
+zstd = ">=0.12, <0.14"
+
+# Internal dependencies
+oxigraph = { version = "=0.4.0-alpha.7-dev", path = "lib/oxigraph" }
+oxrdf = { version = "=0.2.0-alpha.4", path = "lib/oxrdf" }
+oxrdfio = { version = "=0.1.0-alpha.5", path = "lib/oxrdfio" }
+oxrdfxml = { version = "=0.1.0-alpha.5", path = "lib/oxrdfxml" }
+oxsdatatypes = { version = "=0.2.0-alpha.1", path = "lib/oxsdatatypes" }
+oxttl = { version = "=0.1.0-alpha.5", path = "lib/oxttl" }
+sparesults = { version = "=0.2.0-alpha.4", path = "lib/sparesults" }
+spargebra = { version = "=0.3.0-alpha.4", path = "lib/spargebra" }
+sparopt = { version = "=0.1.0-alpha.5-dev", path = "lib/sparopt" }
+
+[workspace.lints.rust]
+absolute_paths_not_starting_with_crate = "warn"
+elided_lifetimes_in_paths = "warn"
+explicit_outlives_requirements = "warn"
+let_underscore_drop = "warn"
+macro_use_extern_crate = "warn"
+# TODO missing_docs = "warn"
+trivial_casts = "warn"
+trivial_numeric_casts = "warn"
+unsafe_code = "warn"
+unused_import_braces = "warn"
+unused_lifetimes = "warn"
+unused_macro_rules = "warn"
+unused_qualifications = "warn"
+
+[workspace.lints.clippy]
+allow_attributes = "warn"
+allow_attributes_without_reason = "warn"
+as_underscore = "warn"
+assertions_on_result_states = "warn"
+bool_to_int_with_if = "warn"
+borrow_as_ptr = "warn"
+case_sensitive_file_extension_comparisons = "warn"
+cast_lossless = "warn"
+cast_possible_truncation = "warn"
+cast_possible_wrap = "warn"
+cast_precision_loss = "warn"
+cast_ptr_alignment = "warn"
+cast_sign_loss = "warn"
+checked_conversions = "warn"
+clone_on_ref_ptr = "warn"
+cloned_instead_of_copied = "warn"
+copy_iterator = "warn"
+create_dir = "warn"
+dbg_macro = "warn"
+decimal_literal_representation = "warn"
+default_trait_access = "warn"
+default_union_representation = "warn"
+deref_by_slicing = "warn"
+disallowed_script_idents = "warn"
+doc_link_with_quotes = "warn"
+empty_drop = "warn"
+empty_enum = "warn"
+empty_structs_with_brackets = "warn"
+enum_glob_use = "warn"
+error_impl_error = "warn"
+exit = "warn"
+expect_used = "warn"
+expl_impl_clone_on_copy = "warn"
+explicit_deref_methods = "warn"
+explicit_into_iter_loop = "warn"
+explicit_iter_loop = "warn"
+filetype_is_file = "warn"
+filter_map_next = "warn"
+flat_map_option = "warn"
+fn_params_excessive_bools = "warn"
+fn_to_numeric_cast_any = "warn"
+format_push_string = "warn"
+from_iter_instead_of_collect = "warn"
+get_unwrap = "warn"
+host_endian_bytes = "warn"
+if_not_else = "warn"
+if_then_some_else_none = "warn"
+ignored_unit_patterns = "warn"
+implicit_clone = "warn"
+implicit_hasher = "warn"
+inconsistent_struct_constructor = "warn"
+index_refutable_slice = "warn"
+inefficient_to_string = "warn"
+infinite_loop = "warn"
+inline_always = "warn"
+inline_asm_x86_att_syntax = "warn"
+inline_asm_x86_intel_syntax = "warn"
+into_iter_without_iter = "warn"
+invalid_upcast_comparisons = "warn"
+items_after_statements = "warn"
+iter_not_returning_iterator = "warn"
+iter_without_into_iter = "warn"
+large_digit_groups = "warn"
+large_futures = "warn"
+large_include_file = "warn"
+large_stack_arrays = "warn"
+large_types_passed_by_value = "warn"
+let_underscore_must_use = "warn"
+let_underscore_untyped = "warn"
+linkedlist = "warn"
+lossy_float_literal = "warn"
+macro_use_imports = "warn"
+manual_assert = "warn"
+manual_instant_elapsed = "warn"
+manual_let_else = "warn"
+manual_ok_or = "warn"
+manual_string_new = "warn"
+many_single_char_names = "warn"
+map_unwrap_or = "warn"
+match_bool = "warn"
+match_on_vec_items = "warn"
+match_same_arms = "warn"
+match_wild_err_arm = "warn"
+match_wildcard_for_single_variants = "warn"
+maybe_infinite_iter = "warn"
+mem_forget = "warn"
+mismatching_type_param_order = "warn"
+missing_assert_message = "warn"
+missing_asserts_for_indexing = "warn"
+missing_fields_in_debug = "warn"
+multiple_inherent_impl = "warn"
+mut_mut = "warn"
+mutex_atomic = "warn"
+naive_bytecount = "warn"
+needless_bitwise_bool = "warn"
+needless_continue = "warn"
+needless_for_each = "warn"
+needless_pass_by_value = "warn"
+needless_raw_string_hashes = "warn"
+needless_raw_strings = "warn"
+negative_feature_names = "warn"
+no_effect_underscore_binding = "warn"
+no_mangle_with_rust_abi = "warn"
+non_ascii_literal = "warn"
+panic = "warn"
+panic_in_result_fn = "warn"
+partial_pub_fields = "warn"
+print_stderr = "warn"
+print_stdout = "warn"
+ptr_as_ptr = "warn"
+ptr_cast_constness = "warn"
+pub_without_shorthand = "warn"
+range_minus_one = "warn"
+range_plus_one = "warn"
+rc_buffer = "warn"
+rc_mutex = "warn"
+redundant_closure_for_method_calls = "warn"
+redundant_else = "warn"
+redundant_feature_names = "warn"
+redundant_type_annotations = "warn"
+ref_binding_to_reference = "warn"
+ref_option_ref = "warn"
+ref_patterns = "warn"
+rest_pat_in_fully_bound_structs = "warn"
+return_self_not_must_use = "warn"
+same_functions_in_if_condition = "warn"
+same_name_method = "warn"
+semicolon_inside_block = "warn"
+shadow_same = "warn"
+should_panic_without_expect = "warn"
+single_match_else = "warn"
+stable_sort_primitive = "warn"
+str_to_string = "warn"
+string_add = "warn"
+string_add_assign = "warn"
+string_lit_chars_any = "warn"
+string_to_string = "warn"
+struct_excessive_bools = "warn"
+struct_field_names = "warn"
+suspicious_xor_used_as_pow = "warn"
+tests_outside_test_module = "warn"
+todo = "warn"
+transmute_ptr_to_ptr = "warn"
+trivially_copy_pass_by_ref = "warn"
+try_err = "warn"
+unchecked_duration_subtraction = "warn"
+undocumented_unsafe_blocks = "warn"
+unicode_not_nfc = "warn"
+unimplemented = "warn"
+uninlined_format_args = "warn"
+unnecessary_box_returns = "warn"
+unnecessary_join = "warn"
+unnecessary_safety_comment = "warn"
+unnecessary_safety_doc = "warn"
+unnecessary_self_imports = "warn"
+unnecessary_wraps = "warn"
+unneeded_field_pattern = "warn"
+unnested_or_patterns = "warn"
+unreadable_literal = "warn"
+unsafe_derive_deserialize = "warn"
+unseparated_literal_suffix = "warn"
+unused_async = "warn"
+unused_self = "warn"
+unwrap_in_result = "warn"
+use_debug = "warn"
+used_underscore_binding = "warn"
+verbose_bit_mask = "warn"
+verbose_file_reads = "warn"
+wildcard_dependencies = "warn"
+zero_sized_map_values = "warn"
+
 [profile.release]
 lto = true
 codegen-units = 1
+strip = "debuginfo"
 
-[profile.release.package.oxigraph_js]
+[profile.release.package.oxigraph-js]
+codegen-units = 1
 opt-level = "z"
+strip = "debuginfo"

README.md

@@ -4,9 +4,10 @@
 [![Released API docs](https://docs.rs/oxigraph/badge.svg)](https://docs.rs/oxigraph)
 [![PyPI](https://img.shields.io/pypi/v/pyoxigraph)](https://pypi.org/project/pyoxigraph/)
 [![npm](https://img.shields.io/npm/v/oxigraph)](https://www.npmjs.com/package/oxigraph)
-[![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
+[![tests status](https://github.com/oxigraph/oxigraph/actions/workflows/tests.yml/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
+[![artifacts status](https://github.com/oxigraph/oxigraph/actions/workflows/artifacts.yml/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
 [![dependency status](https://deps.rs/repo/github/oxigraph/oxigraph/status.svg)](https://deps.rs/repo/github/oxigraph/oxigraph)
-[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge)
+[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community)
 [![Twitter URL](https://img.shields.io/twitter/url?style=social&url=https%3A%2F%2Ftwitter.com%2Foxigraph)](https://twitter.com/oxigraph)
 
 Oxigraph is a graph database implementing the [SPARQL](https://www.w3.org/TR/sparql11-overview/) standard.
@@ -19,7 +20,11 @@ Oxigraph is in heavy development and SPARQL query evaluation has not been optimi
 The development roadmap is using [GitHub milestones](https://github.com/oxigraph/oxigraph/milestones?direction=desc&sort=completeness&state=open).
 Oxigraph internal design [is described on the wiki](https://github.com/oxigraph/oxigraph/wiki/Architecture).
 
-The future 0.4 release is currently in development in the [next branch](https://github.com/oxigraph/oxigraph/tree/next).
+Oxigraph implements the following specifications:
+
+- [SPARQL 1.1 Query](https://www.w3.org/TR/sparql11-query/), [SPARQL 1.1 Update](https://www.w3.org/TR/sparql11-update/), and [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/).
+- [Turtle](https://www.w3.org/TR/turtle/), [TriG](https://www.w3.org/TR/trig/), [N-Triples](https://www.w3.org/TR/n-triples/), [N-Quads](https://www.w3.org/TR/n-quads/), and [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/) RDF serialization formats for both data ingestion and retrieval.
+- [SPARQL Query Results XML Format](https://www.w3.org/TR/rdf-sparql-XMLres/), [SPARQL 1.1 Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/) and [SPARQL 1.1 Query Results CSV and TSV Formats](https://www.w3.org/TR/sparql11-results-csv-tsv/).
 
 It is split into multiple parts:
 
@@ -29,14 +34,22 @@ It is split into multiple parts:
 - [`pyoxigraph` that exposes Oxigraph to the Python world](https://pyoxigraph.readthedocs.io/). Its source code is in the `python` directory. [![PyPI](https://img.shields.io/pypi/v/pyoxigraph)](https://pypi.org/project/pyoxigraph/)
 - [JavaScript bindings for Oxigraph](https://www.npmjs.com/package/oxigraph). WebAssembly is used to package Oxigraph into a NodeJS compatible NPM package. Its source code is in the `js` directory.
   [![npm](https://img.shields.io/npm/v/oxigraph)](https://www.npmjs.com/package/oxigraph)
-- [Oxigraph server](https://crates.io/crates/oxigraph_server) that provides a standalone binary of a web server implementing the [SPARQL 1.1 Protocol](https://www.w3.org/TR/sparql11-protocol/) and the [SPARQL 1.1 Graph Store Protocol](https://www.w3.org/TR/sparql11-http-rdf-update/). Its source code is in the `server` directory.
-  [![Latest Version](https://img.shields.io/crates/v/oxigraph_server.svg)](https://crates.io/crates/oxigraph_server)
-
-Oxigraph implements the following specifications:
-
-- [SPARQL 1.1 Query](https://www.w3.org/TR/sparql11-query/), [SPARQL 1.1 Update](https://www.w3.org/TR/sparql11-update/), and [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/).
-- [Turtle](https://www.w3.org/TR/turtle/), [TriG](https://www.w3.org/TR/trig/), [N-Triples](https://www.w3.org/TR/n-triples/), [N-Quads](https://www.w3.org/TR/n-quads/), and [RDF XML](https://www.w3.org/TR/rdf-syntax-grammar/) RDF serialization formats for both data ingestion and retrieval using the [Rio library](https://github.com/oxigraph/rio).
-- [SPARQL Query Results XML Format](https://www.w3.org/TR/rdf-sparql-XMLres/), [SPARQL 1.1 Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/) and [SPARQL 1.1 Query Results CSV and TSV Formats](https://www.w3.org/TR/sparql11-results-csv-tsv/).
+- [Oxigraph binary](https://crates.io/crates/oxigraph-cli) that provides a standalone command-line tool allowing to manipulate RDF data and spawn a a web server implementing the [SPARQL 1.1 Protocol](https://www.w3.org/TR/sparql11-protocol/) and the [SPARQL 1.1 Graph Store Protocol](https://www.w3.org/TR/sparql11-http-rdf-update/). Its source code is in the `cli` directory.
+  Note that it was previously named [Oxigraph server](https://crates.io/crates/oxigraph-server).
+  [![Latest Version](https://img.shields.io/crates/v/oxigraph-cli.svg)](https://crates.io/crates/oxigraph-cli)
+
+Also, some parts of Oxigraph are available as standalone Rust crates:
+* [`oxrdf`](https://crates.io/crates/oxrdf), datastructures encoding RDF basic concepts (the [`oxigraph::model`](crate::model) module).
+* [`oxrdfio`](https://crates.io/crates/oxrdfio), a unified parser and serializer API for RDF formats (the [`oxigraph::io`](crate::io) module). It itself relies on:
+  * [`oxttl`](https://crates.io/crates/oxttl), N-Triple, N-Quad, Turtle, TriG and N3 parsing and serialization.
+  * [`oxrdfxml`](https://crates.io/crates/oxrdfxml), RDF/XML parsing and serialization.
+* [`spargebra`](https://crates.io/crates/spargebra), a SPARQL parser.
+* [`sparesults`](https://crates.io/crates/sparesults), parsers and serializers for SPARQL result formats.
+* [`sparopt`](https://crates.io/crates/sparesults), a SPARQL optimizer.
+* [`oxsdatatypes`](https://crates.io/crates/oxsdatatypes), an implementation of some XML Schema datatypes.
+
+The library layers in Oxigraph. The elements above depend on the elements below:
+![Oxigraph libraries architecture diagram](./docs/arch-diagram.svg)
 
 A preliminary benchmark [is provided](bench/README.md). There is also [a document describing Oxigraph technical architecture](https://github.com/oxigraph/oxigraph/wiki/Architecture).
 
@@ -72,8 +85,9 @@ Unless you explicitly state otherwise, any contribution intentionally submitted
 ## Sponsors
 
 * [RelationLabs](https://relationlabs.ai/) that is building [Relation-Graph](https://github.com/relationlabs/Relation-Graph), a SPARQL database module for the [Substrate blockchain platform](https://substrate.io/) based on Oxigraph.
-* [Field 33](https://field33.com) that is building [an ontology management platform](https://plow.pm/).
-* [Magnus Bakken](https://github.com/magbak) who is building [chrontext](https://github.com/magbak/chrontext), providing a SPARQL query endpoint on top of joint RDF and time series databases.
+* [Field 33](https://field33.com) that was building [an ontology management platform](https://plow.pm/).
+* [Magnus Bakken](https://github.com/magbak) who is building [Data Treehouse](https://www.data-treehouse.com/), a time-series + RDF datalake platform, and [chrontext](https://github.com/magbak/chrontext), a SPARQL query endpoint on top of joint RDF and time series databases.
+* [DeciSym.AI](https://www.decisym.ai/) a cyber security consulting company providing RDF-based software.
 * [ACE IoT Solutions](https://aceiotsolutions.com/), a building IOT platform.
 * [Albin Larsson](https://byabbe.se/) who is building [GovDirectory](https://www.govdirectory.org/), a directory of public agencies based on Wikidata.
 

bench/bsbm_oxigraph.sh

@@ -6,10 +6,10 @@ PARALLELISM=16
 set -eu
 cd bsbm-tools
 ./generate -fc -pc ${DATASET_SIZE} -s nt -fn "explore-${DATASET_SIZE}" -ud -ufn "explore-update-${DATASET_SIZE}"
-cargo build --release --manifest-path="../../server/Cargo.toml"
-VERSION=$(./../../target/release/oxigraph_server --version | sed 's/oxigraph_server //g')
-./../../target/release/oxigraph_server --location oxigraph_data load --file "explore-${DATASET_SIZE}.nt"
-./../../target/release/oxigraph_server --location oxigraph_data serve --bind 127.0.0.1:7878 &
+cargo build --release --manifest-path="../../cli/Cargo.toml"
+VERSION=$(./../../target/release/oxigraph --version | sed 's/oxigraph //g')
+./../../target/release/oxigraph --location oxigraph_data load --file "explore-${DATASET_SIZE}.nt"
+./../../target/release/oxigraph --location oxigraph_data serve --bind 127.0.0.1:7878 &
 sleep 1
 ./testdriver -mt ${PARALLELISM} -ucf usecases/explore/sparql.txt -o "../bsbm.explore.oxigraph.${VERSION}.${DATASET_SIZE}.${PARALLELISM}.xml" http://127.0.0.1:7878/query
 ./testdriver -mt ${PARALLELISM} -ucf usecases/exploreAndUpdate/sparql.txt -o "../bsbm.exploreAndUpdate.oxigraph.${VERSION}.${DATASET_SIZE}.${PARALLELISM}.xml" http://127.0.0.1:7878/query -u http://127.0.0.1:7878/update -udataset "explore-update-${DATASET_SIZE}.nt"

clippy.toml

@@ -1,4 +1,4 @@
-avoid-breaking-exported-api = true
+avoid-breaking-exported-api = false
 cognitive-complexity-threshold = 50
 too-many-arguments-threshold = 10
 type-complexity-threshold = 500

deny.toml

@@ -1,12 +1,18 @@
+[advisories]
+version = 2
+ignore = ["RUSTSEC-2018-0015"]
+
 [licenses]
-unlicensed = "deny"
+version = 2
 allow = [
-    "MIT",
     "Apache-2.0",
-    "Apache-2.0 WITH LLVM-exception"
+    "Apache-2.0 WITH LLVM-exception",
+    "BSD-3-Clause",
+    "ISC",
+    "MIT",
+    "OpenSSL",
+    "Unicode-DFS-2016"
 ]
-allow-osi-fsf-free = "either"
-default = "deny"
 
 [[licenses.clarify]]
 name = "ring"

docs/arch-diagram.svg

@@ -0,0 +1,120 @@
+<svg xmlns="http://www.w3.org/2000/svg" width="624" height="384" class="svgbob">
+    <style>.svgbob line, .svgbob path, .svgbob circle, .svgbob rect, .svgbob polygon {
+        stroke: black;
+        stroke-width: 2;
+        stroke-opacity: 1;
+        fill-opacity: 1;
+        stroke-linecap: round;
+        stroke-linejoin: miter;
+        }
+        .svgbob text {
+        white-space: pre;
+        fill: black;
+        font-family: Iosevka Fixed, monospace;
+        font-size: 14px;
+        }
+        .svgbob rect.backdrop {
+        stroke: none;
+        fill: white;
+        }
+        .svgbob .broken {
+        stroke-dasharray: 8;
+        }
+        .svgbob .filled {
+        fill: black;
+        }
+        .svgbob .bg_filled {
+        fill: white;
+        stroke-width: 1;
+        }
+        .svgbob .nofill {
+        fill: white;
+        }
+        .svgbob .end_marked_arrow {
+        marker-end: url(#arrow);
+        }
+        .svgbob .start_marked_arrow {
+        marker-start: url(#arrow);
+        }
+        .svgbob .end_marked_diamond {
+        marker-end: url(#diamond);
+        }
+        .svgbob .start_marked_diamond {
+        marker-start: url(#diamond);
+        }
+        .svgbob .end_marked_circle {
+        marker-end: url(#circle);
+        }
+        .svgbob .start_marked_circle {
+        marker-start: url(#circle);
+        }
+        .svgbob .end_marked_open_circle {
+        marker-end: url(#open_circle);
+        }
+        .svgbob .start_marked_open_circle {
+        marker-start: url(#open_circle);
+        }
+        .svgbob .end_marked_big_open_circle {
+        marker-end: url(#big_open_circle);
+        }
+        .svgbob .start_marked_big_open_circle {
+        marker-start: url(#big_open_circle);
+        }<!--separator-->.svgbob .r{
+        fill: papayawhip;
+        }
+        .svgbob .p{
+        fill: lightyellow;
+        }
+        .svgbob .j{
+        fill: lightgreen;
+        }
+    </style>
+    <defs>
+        <marker id="arrow" viewBox="-2 -2 8 8" refX="4" refY="2" markerWidth="7" markerHeight="7"
+                orient="auto-start-reverse">
+            <polygon points="0,0 0,4 4,2 0,0"></polygon>
+        </marker>
+        <marker id="diamond" viewBox="-2 -2 8 8" refX="4" refY="2" markerWidth="7" markerHeight="7"
+                orient="auto-start-reverse">
+            <polygon points="0,2 2,0 4,2 2,4 0,2"></polygon>
+        </marker>
+        <marker id="circle" viewBox="0 0 8 8" refX="4" refY="4" markerWidth="7" markerHeight="7"
+                orient="auto-start-reverse">
+            <circle cx="4" cy="4" r="2" class="filled"></circle>
+        </marker>
+        <marker id="open_circle" viewBox="0 0 8 8" refX="4" refY="4" markerWidth="7" markerHeight="7"
+                orient="auto-start-reverse">
+            <circle cx="4" cy="4" r="2" class="bg_filled"></circle>
+        </marker>
+        <marker id="big_open_circle" viewBox="0 0 8 8" refX="4" refY="4" markerWidth="7" markerHeight="7"
+                orient="auto-start-reverse">
+            <circle cx="4" cy="4" r="3" class="bg_filled"></circle>
+        </marker>
+    </defs>
+    <rect class="backdrop" x="0" y="0" width="624" height="384"></rect>
+    <rect x="4" y="8" width="152" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="18" y="28">oxigraph CLI</text>
+    <rect x="244" y="8" width="136" height="32" class="solid nofill p" rx="0"></rect>
+    <text x="258" y="28">pyoxigraph</text>
+    <rect x="468" y="8" width="144" height="32" class="solid nofill j" rx="0"></rect>
+    <text x="482" y="28">oxigraph</text>
+    <text x="554" y="28">JS</text>
+    <rect x="4" y="72" width="608" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="18" y="92">oxigraph</text>
+    <rect x="68" y="136" width="232" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="82" y="156">oxrdfio</text>
+    <rect x="348" y="136" width="112" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="362" y="156">sparopt</text>
+    <rect x="68" y="200" width="96" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="82" y="220">oxttl</text>
+    <rect x="180" y="200" width="120" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="194" y="220">oxrdfxml</text>
+    <rect x="316" y="200" width="144" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="330" y="220">spargebra</text>
+    <rect x="476" y="200" width="136" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="490" y="220">sparesults</text>
+    <rect x="36" y="264" width="576" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="50" y="284">oxrdf</text>
+    <rect x="4" y="328" width="152" height="32" class="solid nofill r" rx="0"></rect>
+    <text x="18" y="348">oxsdatatypes</text>
+</svg>

docs/arch-diagram.txt

@@ -0,0 +1,35 @@
++------------------+          +----------------+          +-----------------+
++ oxigraph CLI {r} +          + pyoxigraph {p} +          + oxigraph JS {j} +
++------------------+          +----------------+          +-----------------+
+
++---------------------------------------------------------------------------+
++ oxigraph (Rust) {r}                                                       +
++---------------------------------------------------------------------------+
+
+        +----------------------------+     +-------------+
+        + oxrdfio {r}                +     + sparopt {r} +
+        +----------------------------+     +-------------+
+
+        +-----------+ +--------------+ +-----------------+ +----------------+
+        + oxttl {r} + + oxrdfxml {r} + + spargebra {r}   + + sparesults {r} +
+        +-----------+ +--------------+ +-----------------+ +----------------+
+
+    +-----------------------------------------------------------------------+
+    + oxrdf {r}                                                             +
+    +-----------------------------------------------------------------------+
+
++------------------+
++ oxsdatatypes {r} +
++------------------+
+
+
+# Legend:
+r = {
+    fill: papayawhip;
+}
+p = {
+    fill: lightyellow;
+}
+j = {
+    fill: lightgreen;
+}

fuzz/Cargo.toml

@@ -1,7 +1,6 @@
 [package]
 name = "oxigraph-fuzz"
 version = "0.0.0"
-authors = ["Automatically generated"]
 publish = false
 edition = "2021"
 
@@ -9,13 +8,15 @@ edition = "2021"
 cargo-fuzz = true
 
 [dependencies]
-anyhow = "1"
-lazy_static = "1"
+anyhow = "1.0.72"
 libfuzzer-sys = "0.4"
-spargebra = { path = "../lib/spargebra", features = ["rdf-star", "sep-0006"] }
+oxigraph = { path = "../lib/oxigraph", default-features = false }
+oxrdf = { path = "../lib/oxrdf", features = ["rdf-star"] }
+oxrdfxml = { path = "../lib/oxrdfxml" }
+oxttl = { path = "../lib/oxttl", features = ["rdf-star"] }
 sparesults = { path = "../lib/sparesults", features = ["rdf-star"] }
+spargebra = { path = "../lib/spargebra", features = ["rdf-star", "sep-0006"] }
 sparql-smith = { path = "../lib/sparql-smith", features = ["sep-0006"] }
-oxigraph = { path = "../lib" }
 
 [profile.release]
 codegen-units = 1
@@ -23,6 +24,18 @@ debug = true
 
 [workspace]
 
+[[bin]]
+name = "nquads"
+path = "fuzz_targets/nquads.rs"
+
+[[bin]]
+name = "n3"
+path = "fuzz_targets/n3.rs"
+
+[[bin]]
+name = "rdf_xml"
+path = "fuzz_targets/rdf_xml.rs"
+
 [[bin]]
 name = "sparql_eval"
 path = "fuzz_targets/sparql_eval.rs"
@@ -46,3 +59,7 @@ path = "fuzz_targets/sparql_results_xml.rs"
 [[bin]]
 name = "sparql_results_tsv"
 path = "fuzz_targets/sparql_results_tsv.rs"
+
+[[bin]]
+name = "trig"
+path = "fuzz_targets/trig.rs"

fuzz/fuzz_targets/n3.rs

@@ -0,0 +1,28 @@
+#![no_main]
+
+use libfuzzer_sys::fuzz_target;
+use oxttl::N3Parser;
+
+fuzz_target!(|data: &[u8]| {
+    let mut quads = Vec::new();
+    let mut parser = N3Parser::new()
+        .with_base_iri("http://example.com/")
+        .unwrap()
+        .parse();
+    for chunk in data.split(|c| *c == 0xFF) {
+        parser.extend_from_slice(chunk);
+        while let Some(result) = parser.read_next() {
+            if let Ok(quad) = result {
+                quads.push(quad);
+            }
+        }
+    }
+    parser.end();
+    while let Some(result) = parser.read_next() {
+        if let Ok(quad) = result {
+            quads.push(quad);
+        }
+    }
+    assert!(parser.is_end());
+    //TODO: serialize
+});

fuzz/fuzz_targets/nquads.rs

@@ -0,0 +1,84 @@
+#![no_main]
+
+use libfuzzer_sys::fuzz_target;
+use oxrdf::Quad;
+use oxttl::{NQuadsParser, NQuadsSerializer};
+
+fn parse<'a>(
+    chunks: impl IntoIterator<Item = &'a [u8]>,
+    unchecked: bool,
+) -> (Vec<Quad>, Vec<String>) {
+    let mut quads = Vec::new();
+    let mut errors = Vec::new();
+    let mut parser = NQuadsParser::new().with_quoted_triples();
+    if unchecked {
+        parser = parser.unchecked();
+    }
+    let mut reader = parser.parse();
+    for chunk in chunks {
+        reader.extend_from_slice(chunk);
+        while let Some(result) = reader.read_next() {
+            match result {
+                Ok(quad) => quads.push(quad),
+                Err(error) => errors.push(error.to_string()),
+            }
+        }
+    }
+    reader.end();
+    while let Some(result) = reader.read_next() {
+        match result {
+            Ok(quad) => quads.push(quad),
+            Err(error) => errors.push(error.to_string()),
+        }
+    }
+    assert!(reader.is_end());
+    (quads, errors)
+}
+
+fuzz_target!(|data: &[u8]| {
+    // We parse with splitting
+    let (quads, errors) = parse(data.split(|c| *c == 0xFF), false);
+    // We parse without splitting
+    let (quads_without_split, errors_without_split) = parse(
+        [data
+            .iter()
+            .copied()
+            .filter(|c| *c != 0xFF)
+            .collect::<Vec<_>>()
+            .as_slice()],
+        false,
+    );
+    assert_eq!(quads, quads_without_split);
+    assert_eq!(errors, errors_without_split);
+
+    // We test also unchecked if valid
+    if errors.is_empty() {
+        let (quads_unchecked, errors_unchecked) = parse(data.split(|c| *c == 0xFF), true);
+        assert!(errors_unchecked.is_empty());
+        assert_eq!(quads, quads_unchecked);
+    }
+
+    // We serialize
+    let mut writer = NQuadsSerializer::new().serialize_to_write(Vec::new());
+    for quad in &quads {
+        writer.write_quad(quad).unwrap();
+    }
+    let new_serialization = writer.finish();
+
+    // We parse the serialization
+    let new_quads = NQuadsParser::new()
+        .with_quoted_triples()
+        .parse_read(new_serialization.as_slice())
+        .collect::<Result<Vec<_>, _>>()
+        .map_err(|e| {
+            format!(
+                "Error on {:?} from {quads:?} based on {:?}: {e}",
+                String::from_utf8_lossy(&new_serialization),
+                String::from_utf8_lossy(data)
+            )
+        })
+        .unwrap();
+
+    // We check the roundtrip has not changed anything
+    assert_eq!(new_quads, quads);
+});

fuzz/fuzz_targets/rdf_xml.rs

@@ -0,0 +1,35 @@
+#![no_main]
+
+use libfuzzer_sys::fuzz_target;
+use oxrdfxml::{RdfXmlParser, RdfXmlSerializer};
+
+fuzz_target!(|data: &[u8]| {
+    // We parse
+    let triples = RdfXmlParser::new()
+        .parse_read(data)
+        .flatten()
+        .collect::<Vec<_>>();
+
+    // We serialize
+    let mut writer = RdfXmlSerializer::new().serialize_to_write(Vec::new());
+    for triple in &triples {
+        writer.write_triple(triple).unwrap();
+    }
+    let new_serialization = writer.finish().unwrap();
+
+    // We parse the serialization
+    let new_triples = RdfXmlParser::new()
+        .parse_read(new_serialization.as_slice())
+        .collect::<Result<Vec<_>, _>>()
+        .map_err(|e| {
+            format!(
+                "Error on {:?} from {triples:?} based on {:?}: {e}",
+                String::from_utf8_lossy(&new_serialization),
+                String::from_utf8_lossy(data)
+            )
+        })
+        .unwrap();
+
+    // We check the roundtrip has not changed anything
+    assert_eq!(new_triples, triples);
+});

fuzz/fuzz_targets/sparql_eval.rs

@@ -1,31 +1,26 @@
 #![no_main]
 
-use lazy_static::lazy_static;
 use libfuzzer_sys::fuzz_target;
-use oxigraph::io::DatasetFormat;
+use oxigraph::io::RdfFormat;
 use oxigraph::sparql::{Query, QueryOptions, QueryResults, QuerySolutionIter};
 use oxigraph::store::Store;
+use std::sync::OnceLock;
 
-lazy_static! {
-    static ref STORE: Store = {
+fuzz_target!(|data: sparql_smith::Query| {
+    static STORE: OnceLock<Store> = OnceLock::new();
+    let store = STORE.get_or_init(|| {
         let store = Store::new().unwrap();
         store
-            .load_dataset(
-                sparql_smith::DATA_TRIG.as_bytes(),
-                DatasetFormat::TriG,
-                None,
-            )
+            .load_from_read(RdfFormat::TriG, sparql_smith::DATA_TRIG.as_bytes())
             .unwrap();
         store
-    };
-}
+    });
 
-fuzz_target!(|data: sparql_smith::Query| {
     let query_str = data.to_string();
     if let Ok(query) = Query::parse(&query_str, None) {
         let options = QueryOptions::default();
-        let with_opt = STORE.query_opt(query.clone(), options.clone()).unwrap();
-        let without_opt = STORE
+        let with_opt = store.query_opt(query.clone(), options.clone()).unwrap();
+        let without_opt = store
             .query_opt(query, options.without_optimizations())
             .unwrap();
         match (with_opt, without_opt) {
@@ -52,7 +47,7 @@ fn query_solutions_key(iter: QuerySolutionIter, is_reduced: bool) -> String {
             let mut b = t
                 .unwrap()
                 .iter()
-                .map(|(var, val)| format!("{}: {}", var, val))
+                .map(|(var, val)| format!("{var}: {val}"))
                 .collect::<Vec<_>>();
             b.sort_unstable();
             b.join(" ")

fuzz/fuzz_targets/sparql_query.rs

@@ -3,5 +3,5 @@ use libfuzzer_sys::fuzz_target;
 use spargebra::Query;
 
 fuzz_target!(|data: &str| {
-    Query::parse(data, None);
+    let _ = Query::parse(data, None);
 });

fuzz/fuzz_targets/sparql_results_json.rs

@@ -3,4 +3,4 @@ use libfuzzer_sys::fuzz_target;
 use oxigraph_fuzz::result_format::fuzz_result_format;
 use sparesults::QueryResultsFormat;
 
-fuzz_target!(|data: &[u8]| { fuzz_result_format(QueryResultsFormat::Json, data) });
+fuzz_target!(|data: &[u8]| fuzz_result_format(QueryResultsFormat::Json, data));

fuzz/fuzz_targets/sparql_results_tsv.rs

@@ -3,4 +3,4 @@ use libfuzzer_sys::fuzz_target;
 use oxigraph_fuzz::result_format::fuzz_result_format;
 use sparesults::QueryResultsFormat;
 
-fuzz_target!(|data: &[u8]| { fuzz_result_format(QueryResultsFormat::Tsv, data) });
+fuzz_target!(|data: &[u8]| fuzz_result_format(QueryResultsFormat::Tsv, data));

fuzz/fuzz_targets/sparql_results_xml.rs

@@ -3,4 +3,4 @@ use libfuzzer_sys::fuzz_target;
 use oxigraph_fuzz::result_format::fuzz_result_format;
 use sparesults::QueryResultsFormat;
 
-fuzz_target!(|data: &[u8]| { fuzz_result_format(QueryResultsFormat::Xml, data) });
+fuzz_target!(|data: &[u8]| fuzz_result_format(QueryResultsFormat::Xml, data));

fuzz/fuzz_targets/sparql_update.rs

@@ -4,5 +4,5 @@ use spargebra::Update;
 use std::str;
 
 fuzz_target!(|data: &str| {
-    Update::parse(data, None);
+    let _ = Update::parse(data, None);
 });

fuzz/fuzz_targets/trig.rs

@@ -0,0 +1,166 @@
+#![no_main]
+
+use libfuzzer_sys::fuzz_target;
+use oxrdf::graph::CanonicalizationAlgorithm;
+use oxrdf::{Dataset, GraphName, Quad, Subject, Term, Triple};
+use oxttl::{TriGParser, TriGSerializer};
+
+fn parse<'a>(
+    chunks: impl IntoIterator<Item = &'a [u8]>,
+    unchecked: bool,
+) -> (Vec<Quad>, Vec<String>, Vec<(String, String)>) {
+    let mut quads = Vec::new();
+    let mut errors = Vec::new();
+    let mut parser = TriGParser::new()
+        .with_quoted_triples()
+        .with_base_iri("http://example.com/")
+        .unwrap();
+    if unchecked {
+        parser = parser.unchecked();
+    }
+    let mut reader = parser.parse();
+    for chunk in chunks {
+        reader.extend_from_slice(chunk);
+        while let Some(result) = reader.read_next() {
+            match result {
+                Ok(quad) => quads.push(quad),
+                Err(error) => errors.push(error.to_string()),
+            }
+        }
+    }
+    reader.end();
+    while let Some(result) = reader.read_next() {
+        match result {
+            Ok(quad) => quads.push(quad),
+            Err(error) => errors.push(error.to_string()),
+        }
+    }
+    assert!(reader.is_end());
+    (
+        quads,
+        errors,
+        reader
+            .prefixes()
+            .map(|(k, v)| (k.to_owned(), v.to_owned()))
+            .collect(),
+    )
+}
+
+fn count_triple_blank_nodes(triple: &Triple) -> usize {
+    (match &triple.subject {
+        Subject::BlankNode(_) => 1,
+        Subject::Triple(t) => count_triple_blank_nodes(t),
+        _ => 0,
+    }) + (match &triple.object {
+        Term::BlankNode(_) => 1,
+        Term::Triple(t) => count_triple_blank_nodes(t),
+        _ => 0,
+    })
+}
+
+fn count_quad_blank_nodes(quad: &Quad) -> usize {
+    (match &quad.subject {
+        Subject::BlankNode(_) => 1,
+        Subject::Triple(t) => count_triple_blank_nodes(t),
+        _ => 0,
+    }) + (match &quad.object {
+        Term::BlankNode(_) => 1,
+        Term::Triple(t) => count_triple_blank_nodes(t),
+        _ => 0,
+    }) + usize::from(matches!(quad.graph_name, GraphName::BlankNode(_)))
+}
+
+fn serialize_quads(quads: &[Quad], prefixes: Vec<(String, String)>) -> Vec<u8> {
+    let mut serializer = TriGSerializer::new();
+    for (prefix_name, prefix_iri) in prefixes {
+        serializer = serializer.with_prefix(prefix_name, prefix_iri).unwrap();
+    }
+    let mut writer = serializer.serialize_to_write(Vec::new());
+    for quad in quads {
+        writer.write_quad(quad).unwrap();
+    }
+    writer.finish().unwrap()
+}
+
+fuzz_target!(|data: &[u8]| {
+    // We parse with splitting
+    let (quads, errors, prefixes) = parse(data.split(|c| *c == 0xFF), false);
+    // We parse without splitting
+    let (quads_without_split, errors_without_split, _) = parse(
+        [data
+            .iter()
+            .copied()
+            .filter(|c| *c != 0xFF)
+            .collect::<Vec<_>>()
+            .as_slice()],
+        false,
+    );
+    let (quads_unchecked, errors_unchecked, _) = parse(data.split(|c| *c == 0xFF), true);
+    if errors.is_empty() {
+        assert!(errors_unchecked.is_empty());
+    }
+
+    let bnodes_count = quads.iter().map(count_quad_blank_nodes).sum::<usize>();
+    if bnodes_count == 0 {
+        assert_eq!(
+            quads,
+            quads_without_split,
+            "With split:\n{}\nWithout split:\n{}",
+            String::from_utf8_lossy(&serialize_quads(&quads, Vec::new())),
+            String::from_utf8_lossy(&serialize_quads(&quads_without_split, Vec::new()))
+        );
+        if errors.is_empty() {
+            assert_eq!(
+                quads,
+                quads_unchecked,
+                "Validating:\n{}\nUnchecked:\n{}",
+                String::from_utf8_lossy(&serialize_quads(&quads, Vec::new())),
+                String::from_utf8_lossy(&serialize_quads(&quads_unchecked, Vec::new()))
+            );
+        }
+    } else if bnodes_count <= 4 {
+        let mut dataset_with_split = quads.iter().collect::<Dataset>();
+        let mut dataset_without_split = quads_without_split.iter().collect::<Dataset>();
+        dataset_with_split.canonicalize(CanonicalizationAlgorithm::Unstable);
+        dataset_without_split.canonicalize(CanonicalizationAlgorithm::Unstable);
+        assert_eq!(
+            dataset_with_split,
+            dataset_without_split,
+            "With split:\n{}\nWithout split:\n{}",
+            String::from_utf8_lossy(&serialize_quads(&quads, Vec::new())),
+            String::from_utf8_lossy(&serialize_quads(&quads_without_split, Vec::new()))
+        );
+        if errors.is_empty() {
+            let mut dataset_unchecked = quads_unchecked.iter().collect::<Dataset>();
+            dataset_unchecked.canonicalize(CanonicalizationAlgorithm::Unstable);
+            assert_eq!(
+                dataset_with_split,
+                dataset_unchecked,
+                "Validating:\n{}\nUnchecked:\n{}",
+                String::from_utf8_lossy(&serialize_quads(&quads, Vec::new())),
+                String::from_utf8_lossy(&serialize_quads(&quads_unchecked, Vec::new()))
+            );
+        }
+    }
+    assert_eq!(errors, errors_without_split);
+
+    // We serialize
+    let new_serialization = serialize_quads(&quads, prefixes);
+
+    // We parse the serialization
+    let new_quads = TriGParser::new()
+        .with_quoted_triples()
+        .parse_read(new_serialization.as_slice())
+        .collect::<Result<Vec<_>, _>>()
+        .map_err(|e| {
+            format!(
+                "Error on {:?} from {quads:?} based on {:?}: {e}",
+                String::from_utf8_lossy(&new_serialization),
+                String::from_utf8_lossy(data)
+            )
+        })
+        .unwrap();
+
+    // We check the roundtrip has not changed anything
+    assert_eq!(new_quads, quads);
+});

fuzz/src/result_format.rs

@@ -1,27 +1,27 @@
 use anyhow::Context;
 use sparesults::{
-    QueryResultsFormat, QueryResultsParser, QueryResultsReader, QueryResultsSerializer,
+    FromReadQueryResultsReader, QueryResultsFormat, QueryResultsParser, QueryResultsSerializer,
 };
 
 pub fn fuzz_result_format(format: QueryResultsFormat, data: &[u8]) {
     let parser = QueryResultsParser::from_format(format);
     let serializer = QueryResultsSerializer::from_format(format);
 
-    let Ok(reader) = parser.read_results(data) else {
+    let Ok(reader) = parser.parse_read(data) else {
         return;
     };
     match reader {
-        QueryResultsReader::Solutions(solutions) => {
+        FromReadQueryResultsReader::Solutions(solutions) => {
             let Ok(solutions) = solutions.collect::<Result<Vec<_>, _>>() else {
                 return;
             };
 
             // We try to write again
             let mut writer = serializer
-                .solutions_writer(
+                .serialize_solutions_to_write(
                     Vec::new(),
                     solutions
-                        .get(0)
+                        .first()
                         .map_or_else(Vec::new, |s| s.variables().to_vec()),
                 )
                 .unwrap();
@@ -31,30 +31,30 @@ pub fn fuzz_result_format(format: QueryResultsFormat, data: &[u8]) {
             let serialized = String::from_utf8(writer.finish().unwrap()).unwrap();
 
             // And to parse again
-            if let QueryResultsReader::Solutions(roundtrip_solutions) = parser
-                .read_results(serialized.as_bytes())
-                .with_context(|| format!("Parsing {:?}", &serialized))
+            if let FromReadQueryResultsReader::Solutions(roundtrip_solutions) = parser
+                .parse_read(serialized.as_bytes())
+                .with_context(|| format!("Parsing {serialized:?}"))
                 .unwrap()
             {
                 assert_eq!(
                     roundtrip_solutions
                         .collect::<Result<Vec<_>, _>>()
-                        .with_context(|| format!("Parsing {:?}", &serialized))
+                        .with_context(|| format!("Parsing {serialized:?}"))
                         .unwrap(),
                     solutions
                 )
             }
         }
-        QueryResultsReader::Boolean(value) => {
+        FromReadQueryResultsReader::Boolean(value) => {
             // We try to write again
             let mut serialized = Vec::new();
             serializer
-                .write_boolean_result(&mut serialized, value)
+                .serialize_boolean_to_write(&mut serialized, value)
                 .unwrap();
 
             // And to parse again
-            if let QueryResultsReader::Boolean(roundtrip_value) =
-                parser.read_results(serialized.as_slice()).unwrap()
+            if let FromReadQueryResultsReader::Boolean(roundtrip_value) =
+                parser.parse_read(serialized.as_slice()).unwrap()
             {
                 assert_eq!(roundtrip_value, value)
             }

js/Cargo.toml

@@ -1,20 +1,26 @@
 [package]
-name = "oxigraph_js"
-version = "0.3.22"
-authors = ["Tpt <thomas@pellissier-tanon.fr>"]
-license = "MIT OR Apache-2.0"
+name = "oxigraph-js"
+version.workspace = true
+authors.workspace = true
+license.workspace = true
 readme = "README.md"
-keywords = ["RDF", "N-Triples", "Turtle", "RDF/XML", "SPARQL"]
+keywords = ["RDF", "N-Triples", "Turtle", "XML", "SPARQL"]
 repository = "https://github.com/oxigraph/oxigraph/tree/main/js"
 description = "JavaScript bindings of Oxigraph"
-edition = "2021"
+edition.workspace = true
+rust-version.workspace = true
+publish = false
 
 [lib]
 crate-type = ["cdylib"]
 name = "oxigraph"
+doc = false
 
 [dependencies]
-oxigraph = { version = "0.3.22", path="../lib" }
-wasm-bindgen = "0.2"
-js-sys = "0.3"
-console_error_panic_hook = "0.1"
+console_error_panic_hook.workspace = true
+js-sys.workspace = true
+oxigraph = { workspace = true, features = ["js"] }
+wasm-bindgen.workspace = true
+
+[lints]
+workspace = true

js/README.md

@@ -3,7 +3,7 @@ Oxigraph for JavaScript
 
 [![npm](https://img.shields.io/npm/v/oxigraph)](https://www.npmjs.com/package/oxigraph)
 [![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
-[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge)
+[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community)
 
 This package provides a JavaScript API on top of [Oxigraph](https://crates.io/crates/oxigraph), compiled with WebAssembly.
 
@@ -13,7 +13,7 @@ Oxigraph for JavaScript is a work in progress and currently offers a simple in-m
 
 The store is also able to load RDF serialized in [Turtle](https://www.w3.org/TR/turtle/), [TriG](https://www.w3.org/TR/trig/), [N-Triples](https://www.w3.org/TR/n-triples/), [N-Quads](https://www.w3.org/TR/n-quads/) and [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/).
 
-It is distributed using a [a NPM package](https://www.npmjs.com/package/oxigraph) that should work with Node.JS 12+ and modern web browsers compatible with WebAssembly.
+It is distributed using a [a NPM package](https://www.npmjs.com/package/oxigraph) that should work with Node.JS 18+ and [modern web browsers compatible with WebAssembly reference types and JavaScript `WeakRef`](https://caniuse.com/wasm-reference-types,mdn-javascript_builtins_weakref).
 
 To install:
 ```bash
@@ -188,6 +188,18 @@ if (store.query("ASK { ?s ?s ?s }")) {
 }
 ```
 
+It is also possible to provide some options in an object given as second argument:
+
+```js
+if (store.query("ASK { <s> ?p ?o }", {
+  base_iri: "http://example.com/", // base IRI to resolve relative IRIs in the query
+  use_default_graph_as_union: true, // the default graph in the query is the union of all the dataset graphs
+
+})) {
+  console.log("there is a triple with same subject, predicate and object");
+}
+```
+
 #### `Store.prototype.update(String query)`
 Executes a [SPARQL 1.1 Update](https://www.w3.org/TR/sparql11-update/).
 The [`LOAD` operation](https://www.w3.org/TR/sparql11-update/#load) is not supported yet.
@@ -197,40 +209,50 @@ Example of update:
 store.update("DELETE WHERE { <http://example.com/s> ?p ?o }")
 ```
 
-#### `Store.prototype.load(String data, String mimeType, NamedNode|String? baseIRI, NamedNode|BlankNode|DefaultGraph? toNamedGraph)`
+It is also possible to provide some options in an object given as second argument:
+
+```js
+store.update("DELETE WHERE { <s> ?p ?o }", {
+  base_iri: "http://example.com/" // base IRI to resolve relative IRIs in the update
+})
+```
+
+#### `Store.prototype.load(String data, String format, NamedNode|String? baseIRI, NamedNode|BlankNode|DefaultGraph? toNamedGraph)`
 
 Loads serialized RDF triples or quad into the store.
 The method arguments are:
 1. `data`: the serialized RDF triples or quads.
-2. `mimeType`: the MIME type of the serialization. See below for the supported mime types.
+2. `format`: the format of the serialization. See below for the supported formats.
 3. `baseIRI`: the base IRI to use to resolve the relative IRIs in the serialization.
 4. `toNamedGraph`: for triple serialization formats, the name of the named graph the triple should be loaded to.
 
 The available formats are:
-* [Turtle](https://www.w3.org/TR/turtle/): `text/turtle`
-* [TriG](https://www.w3.org/TR/trig/): `application/trig`
-* [N-Triples](https://www.w3.org/TR/n-triples/): `application/n-triples`
-* [N-Quads](https://www.w3.org/TR/n-quads/): `application/n-quads`
-* [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/): `application/rdf+xml`
+* [Turtle](https://www.w3.org/TR/turtle/): `text/turtle` or `ttl`
+* [TriG](https://www.w3.org/TR/trig/): `application/trig` or `trig`
+* [N-Triples](https://www.w3.org/TR/n-triples/): `application/n-triples` or `nt`
+* [N-Quads](https://www.w3.org/TR/n-quads/): `application/n-quads` or `nq`
+* [N3](https://w3c.github.io/N3/spec/): `text/n3` or `n3`
+* [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/): `application/rdf+xml` or `rdf`
 
 Example of loading a Turtle file into the named graph `<http://example.com/graph>` with the base IRI `http://example.com`:
 ```js
 store.load("<http://example.com> <http://example.com> <> .", "text/turtle", "http://example.com", oxigraph.namedNode("http://example.com/graph"));
 ```
 
-#### `Store.prototype.dump(String mimeType, NamedNode|BlankNode|DefaultGraph? fromNamedGraph)`
+#### `Store.prototype.dump(String format, NamedNode|BlankNode|DefaultGraph? fromNamedGraph)`
 
 Returns serialized RDF triples or quad from the store.
 The method arguments are:
-1. `mimeType`: the MIME type of the serialization. See below for the supported mime types.
+1. `format`: the format type of the serialization. See below for the supported types.
 2. `fromNamedGraph`: for triple serialization formats, the name of the named graph the triple should be loaded from.
 
 The available formats are:
-* [Turtle](https://www.w3.org/TR/turtle/): `text/turtle`
-* [TriG](https://www.w3.org/TR/trig/): `application/trig`
-* [N-Triples](https://www.w3.org/TR/n-triples/): `application/n-triples`
-* [N-Quads](https://www.w3.org/TR/n-quads/): `application/n-quads`
-* [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/): `application/rdf+xml`
+* [Turtle](https://www.w3.org/TR/turtle/): `text/turtle` or `ttl`
+* [TriG](https://www.w3.org/TR/trig/): `application/trig` or `trig`
+* [N-Triples](https://www.w3.org/TR/n-triples/): `application/n-triples` or `nt`
+* [N-Quads](https://www.w3.org/TR/n-quads/): `application/n-quads` or `nq`
+* [N3](https://w3c.github.io/N3/spec/): `text/n3` or `n3`
+* [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/): `application/rdf+xml` or `rdf`
 
 Example of building a Turtle file from the named graph `<http://example.com/graph>`:
 ```js
@@ -260,7 +282,7 @@ To setup a dev environment:
 Testing and linting:
 - Rust code is formatted with [rustfmt](https://github.com/rust-lang/rustfmt) and linted with [clippy](https://github.com/rust-lang/rust-clippy).
   You can execute them with `cargo fmt` and `cargo clippy`.
-- JS code is formatted and linted with [Rome](https://rome.tools/). `npm run fmt` to auto-format and `npm test` to lint and test.
+- JS code is formatted and linted with [Biome](https://biomejs.dev/). `npm run fmt` to auto-format and `npm test` to lint and test.
 - Tests are written in JavaScript using [Mocha](https://mochajs.org/) in the `test` directory. `npm test` to run them.
 
 

js/biome.json

@@ -1,7 +1,8 @@
 {
+    "$schema": "https://biomejs.dev/schemas/1.0.0/schema.json",
     "formatter": {
         "indentStyle": "space",
-        "indentSize": 4,
+        "indentWidth": 4,
         "lineWidth": 100
     },
     "linter": {

js/build_package.js

@@ -1,16 +1,6 @@
 #! /usr/bin/env node
 
-const fs = require("fs");
-
-// We copy file to the new directory
-fs.mkdirSync("pkg");
-for (const file of fs.readdirSync("./pkg-web")) {
-    fs.copyFileSync(`./pkg-web/${file}`, `./pkg/${file}`);
-}
-for (const file of fs.readdirSync("./pkg-node")) {
-    fs.copyFileSync(`./pkg-node/${file}`, `./pkg/${file}`);
-}
-
+const fs = require("node:fs");
 const pkg = JSON.parse(fs.readFileSync("./pkg/package.json"));
 pkg.name = "oxigraph";
 pkg.main = "node.js";

js/package.json

@@ -1,22 +1,20 @@
 {
-  "name": "oxigraph_tests",
-  "description": "Oxigraph JS build and tests",
-  "private": true,
-  "devDependencies": {
-    "@rdfjs/data-model": "^2.0.1",
-    "mocha": "^10.0.0",
-    "rome": "^12.0.0"
-  },
-  "scripts": {
-    "fmt": "rome format . --write && rome check . --apply-unsafe",
-    "test": "rome ci . && wasm-pack build --debug --target nodejs && mocha",
-    "build": "rm -rf pkg && wasm-pack build --release --target web --out-name web && mv pkg pkg-web && wasm-pack build --release --target nodejs --out-name node && mv pkg pkg-node && node build_package.js && rm -r pkg-web && rm -r pkg-node",
-    "release": "npm run build && npm publish ./pkg",
-    "pack": "npm run build && npm pack ./pkg"
-  },
-  "standard": {
-    "ignore": [
-      "pkg*"
-    ]
-  }
+    "name": "oxigraph_tests",
+    "description": "Oxigraph JS build and tests",
+    "private": true,
+    "devDependencies": {
+        "@biomejs/biome": "^1.0.0",
+        "@rdfjs/data-model": "^2.0.1",
+        "mocha": "^10.0.0"
+    },
+    "scripts": {
+        "fmt": "biome format . --write && biome check . --apply-unsafe && biome format . --write",
+        "test": "biome ci . && wasm-pack build --debug --target nodejs --weak-refs --reference-types && mocha",
+        "build": "wasm-pack build --release --target web --out-name web --weak-refs --reference-types && wasm-pack build --release --target nodejs --out-name node --weak-refs --reference-types && node build_package.js",
+        "release": "npm run build && npm publish ./pkg",
+        "pack": "npm run build && npm pack ./pkg"
+    },
+    "standard": {
+        "ignore": ["pkg*"]
+    }
 }

js/src/lib.rs

@@ -1,4 +1,4 @@
-#![allow(clippy::unused_unit)]
+#![allow(clippy::mem_forget)]
 use wasm_bindgen::prelude::*;
 
 mod model;

js/src/model.rs

@@ -1,15 +1,9 @@
-#![allow(
-    dead_code,
-    clippy::inherent_to_string,
-    clippy::unused_self,
-    clippy::use_self
-)]
+#![allow(dead_code, clippy::inherent_to_string, clippy::unused_self)]
 
 use crate::format_err;
 use crate::utils::to_err;
 use js_sys::{Reflect, UriError};
 use oxigraph::model::*;
-use oxigraph::sparql::Variable;
 use wasm_bindgen::prelude::*;
 
 thread_local! {
@@ -56,7 +50,7 @@ pub fn literal(
 
 #[wasm_bindgen(js_name = defaultGraph)]
 pub fn default_graph() -> JsDefaultGraph {
-    JsDefaultGraph {}
+    JsDefaultGraph
 }
 
 #[wasm_bindgen(js_name = variable)]
@@ -302,7 +296,7 @@ impl From<JsLiteral> for Term {
 
 #[wasm_bindgen(js_name = DefaultGraph)]
 #[derive(Eq, PartialEq, Debug, Clone, Hash)]
-pub struct JsDefaultGraph {}
+pub struct JsDefaultGraph;
 
 #[wasm_bindgen(js_class = DefaultGraph)]
 impl JsDefaultGraph {
@@ -313,7 +307,7 @@ impl JsDefaultGraph {
 
     #[wasm_bindgen(getter)]
     pub fn value(&self) -> String {
-        "".to_owned()
+        String::new()
     }
 
     #[wasm_bindgen(js_name = toString)]
@@ -393,7 +387,7 @@ impl JsQuad {
 
     #[wasm_bindgen(getter)]
     pub fn value(&self) -> String {
-        "".to_owned()
+        String::new()
     }
 
     #[wasm_bindgen(getter = subject)]
@@ -532,7 +526,7 @@ impl From<GraphName> for JsTerm {
         match name {
             GraphName::NamedNode(node) => node.into(),
             GraphName::BlankNode(node) => node.into(),
-            GraphName::DefaultGraph => Self::DefaultGraph(JsDefaultGraph {}),
+            GraphName::DefaultGraph => Self::DefaultGraph(JsDefaultGraph),
         }
     }
 }
@@ -564,7 +558,7 @@ impl From<Quad> for JsTerm {
 impl TryFrom<JsTerm> for NamedNode {
     type Error = JsValue;
 
-    fn try_from(value: JsTerm) -> Result<Self, JsValue> {
+    fn try_from(value: JsTerm) -> Result<Self, Self::Error> {
         match value {
             JsTerm::NamedNode(node) => Ok(node.into()),
             JsTerm::BlankNode(node) => Err(format_err!(
@@ -588,7 +582,7 @@ impl TryFrom<JsTerm> for NamedNode {
 impl TryFrom<JsTerm> for NamedOrBlankNode {
     type Error = JsValue;
 
-    fn try_from(value: JsTerm) -> Result<Self, JsValue> {
+    fn try_from(value: JsTerm) -> Result<Self, Self::Error> {
         match value {
             JsTerm::NamedNode(node) => Ok(node.into()),
             JsTerm::BlankNode(node) => Ok(node.into()),
@@ -614,7 +608,7 @@ impl TryFrom<JsTerm> for NamedOrBlankNode {
 impl TryFrom<JsTerm> for Subject {
     type Error = JsValue;
 
-    fn try_from(value: JsTerm) -> Result<Self, JsValue> {
+    fn try_from(value: JsTerm) -> Result<Self, Self::Error> {
         match value {
             JsTerm::NamedNode(node) => Ok(node.into()),
             JsTerm::BlankNode(node) => Ok(node.into()),
@@ -637,7 +631,7 @@ impl TryFrom<JsTerm> for Subject {
 impl TryFrom<JsTerm> for Term {
     type Error = JsValue;
 
-    fn try_from(value: JsTerm) -> Result<Self, JsValue> {
+    fn try_from(value: JsTerm) -> Result<Self, Self::Error> {
         match value {
             JsTerm::NamedNode(node) => Ok(node.into()),
             JsTerm::BlankNode(node) => Ok(node.into()),
@@ -657,7 +651,7 @@ impl TryFrom<JsTerm> for Term {
 impl TryFrom<JsTerm> for GraphName {
     type Error = JsValue;
 
-    fn try_from(value: JsTerm) -> Result<Self, JsValue> {
+    fn try_from(value: JsTerm) -> Result<Self, Self::Error> {
         match value {
             JsTerm::NamedNode(node) => Ok(node.into()),
             JsTerm::BlankNode(node) => Ok(node.into()),
@@ -744,7 +738,7 @@ impl FromJsConverter {
                         ))
                     }
                 }
-                "DefaultGraph" => Ok(JsTerm::DefaultGraph(JsDefaultGraph {})),
+                "DefaultGraph" => Ok(JsTerm::DefaultGraph(JsDefaultGraph)),
                 "Variable" => Ok(Variable::new(
                     Reflect::get(value, &self.value)?
                         .as_string()
@@ -754,8 +748,7 @@ impl FromJsConverter {
                 .into()),
                 "Quad" => Ok(self.to_quad(value)?.into()),
                 _ => Err(format_err!(
-                    "The termType {} is not supported by Oxigraph",
-                    term_type
+                    "The termType {term_type} is not supported by Oxigraph"
                 )),
             }
         } else if term_type.is_undefined() {

js/src/store.rs

@@ -1,14 +1,11 @@
-#![allow(clippy::use_self)]
-
 use crate::format_err;
 use crate::model::*;
 use crate::utils::to_err;
-use js_sys::{Array, Map};
-use oxigraph::io::{DatasetFormat, GraphFormat};
+use js_sys::{Array, Map, Reflect};
+use oxigraph::io::{RdfFormat, RdfParser};
 use oxigraph::model::*;
-use oxigraph::sparql::QueryResults;
+use oxigraph::sparql::{Query, QueryResults, Update};
 use oxigraph::store::Store;
-use std::io::Cursor;
 use wasm_bindgen::prelude::*;
 
 #[wasm_bindgen(js_name = Store)]
@@ -27,7 +24,7 @@ impl JsStore {
             store: Store::new().map_err(to_err)?,
         };
         if let Some(quads) = quads {
-            for quad in quads.iter() {
+            for quad in &*quads {
                 store.add(quad)?;
             }
         }
@@ -105,7 +102,21 @@ impl JsStore {
             .into_boxed_slice())
     }
 
-    pub fn query(&self, query: &str) -> Result<JsValue, JsValue> {
+    pub fn query(&self, query: &str, options: &JsValue) -> Result<JsValue, JsValue> {
+        // Parsing options
+        let mut base_iri = None;
+        let mut use_default_graph_as_union = false;
+        if !options.is_undefined() {
+            base_iri = Reflect::get(options, &JsValue::from_str("base_iri"))?.as_string();
+            use_default_graph_as_union =
+                Reflect::get(options, &JsValue::from_str("use_default_graph_as_union"))?
+                    .is_truthy();
+        }
+
+        let mut query = Query::parse(query, base_iri.as_deref()).map_err(to_err)?;
+        if use_default_graph_as_union {
+            query.dataset_mut().set_default_graph_as_union();
+        }
         let results = self.store.query(query).map_err(to_err)?;
         let output = match results {
             QueryResults::Solutions(solutions) => {
@@ -138,17 +149,25 @@ impl JsStore {
         Ok(output)
     }
 
-    pub fn update(&self, update: &str) -> Result<(), JsValue> {
+    pub fn update(&self, update: &str, options: &JsValue) -> Result<(), JsValue> {
+        // Parsing options
+        let mut base_iri = None;
+        if !options.is_undefined() {
+            base_iri = Reflect::get(options, &JsValue::from_str("base_iri"))?.as_string();
+        }
+
+        let update = Update::parse(update, base_iri.as_deref()).map_err(to_err)?;
         self.store.update(update).map_err(to_err)
     }
 
     pub fn load(
         &self,
         data: &str,
-        mime_type: &str,
+        format: &str,
         base_iri: &JsValue,
         to_graph_name: &JsValue,
     ) -> Result<(), JsValue> {
+        let format = rdf_format(format)?;
         let base_iri = if base_iri.is_null() || base_iri.is_undefined() {
             None
         } else if base_iri.is_string() {
@@ -161,65 +180,41 @@ impl JsStore {
             ));
         };
 
-        let to_graph_name =
-            if let Some(graph_name) = FROM_JS.with(|c| c.to_optional_term(to_graph_name))? {
-                Some(graph_name.try_into()?)
-            } else {
-                None
-            };
-
-        if let Some(graph_format) = GraphFormat::from_media_type(mime_type) {
-            self.store
-                .load_graph(
-                    Cursor::new(data),
-                    graph_format,
-                    &to_graph_name.unwrap_or(GraphName::DefaultGraph),
-                    base_iri.as_deref(),
-                )
-                .map_err(to_err)
-        } else if let Some(dataset_format) = DatasetFormat::from_media_type(mime_type) {
-            if to_graph_name.is_some() {
-                return Err(format_err!(
-                    "The target graph name parameter is not available for dataset formats"
-                ));
-            }
-            self.store
-                .load_dataset(Cursor::new(data), dataset_format, base_iri.as_deref())
-                .map_err(to_err)
-        } else {
-            Err(format_err!("Not supported MIME type: {}", mime_type))
+        let mut parser = RdfParser::from_format(format);
+        if let Some(to_graph_name) = FROM_JS.with(|c| c.to_optional_term(to_graph_name))? {
+            parser = parser.with_default_graph(GraphName::try_from(to_graph_name)?);
+        }
+        if let Some(base_iri) = base_iri {
+            parser = parser.with_base_iri(base_iri).map_err(to_err)?;
         }
+        self.store
+            .load_from_read(parser, data.as_bytes())
+            .map_err(to_err)
     }
 
-    pub fn dump(&self, mime_type: &str, from_graph_name: &JsValue) -> Result<String, JsValue> {
-        let from_graph_name =
-            if let Some(graph_name) = FROM_JS.with(|c| c.to_optional_term(from_graph_name))? {
-                Some(graph_name.try_into()?)
-            } else {
-                None
-            };
-
-        let mut buffer = Vec::new();
-        if let Some(graph_format) = GraphFormat::from_media_type(mime_type) {
-            self.store
-                .dump_graph(
-                    &mut buffer,
-                    graph_format,
-                    &from_graph_name.unwrap_or(GraphName::DefaultGraph),
+    pub fn dump(&self, format: &str, from_graph_name: &JsValue) -> Result<String, JsValue> {
+        let format = rdf_format(format)?;
+        let buffer =
+            if let Some(from_graph_name) = FROM_JS.with(|c| c.to_optional_term(from_graph_name))? {
+                self.store.dump_graph_to_write(
+                    &GraphName::try_from(from_graph_name)?,
+                    format,
+                    Vec::new(),
                 )
-                .map_err(to_err)?;
-        } else if let Some(dataset_format) = DatasetFormat::from_media_type(mime_type) {
-            if from_graph_name.is_some() {
-                return Err(format_err!(
-                    "The target graph name parameter is not available for dataset formats"
-                ));
+            } else {
+                self.store.dump_to_write(format, Vec::new())
             }
-            self.store
-                .dump_dataset(&mut buffer, dataset_format)
-                .map_err(to_err)?;
-        } else {
-            return Err(format_err!("Not supported MIME type: {}", mime_type));
-        }
+            .map_err(to_err)?;
         String::from_utf8(buffer).map_err(to_err)
     }
 }
+
+fn rdf_format(format: &str) -> Result<RdfFormat, JsValue> {
+    if format.contains('/') {
+        RdfFormat::from_media_type(format)
+            .ok_or_else(|| format_err!("Not supported RDF format media type: {format}"))
+    } else {
+        RdfFormat::from_extension(format)
+            .ok_or_else(|| format_err!("Not supported RDF format extension: {format}"))
+    }
+}

js/test/model.mjs

@@ -1,37 +1,37 @@
 /* global describe, it */
 
+import assert from "node:assert";
 import runTests from "../node_modules/@rdfjs/data-model/test/index.js";
 import oxigraph from "../pkg/oxigraph.js";
-import assert from "assert";
 
 runTests({ factory: oxigraph });
 
-describe("DataModel", function () {
-    describe("#toString()", function () {
-        it("namedNode().toString() should return SPARQL compatible syntax", function () {
+describe("DataModel", () => {
+    describe("#toString()", () => {
+        it("namedNode().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual(
                 "<http://example.com>",
                 oxigraph.namedNode("http://example.com").toString(),
             );
         });
 
-        it("blankNode().toString() should return SPARQL compatible syntax", function () {
+        it("blankNode().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual("_:a", oxigraph.blankNode("a").toString());
         });
 
-        it("literal().toString() should return SPARQL compatible syntax", function () {
+        it("literal().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual('"a\\"b"@en', oxigraph.literal('a"b', "en").toString());
         });
 
-        it("defaultGraph().toString() should return SPARQL compatible syntax", function () {
+        it("defaultGraph().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual("DEFAULT", oxigraph.defaultGraph().toString());
         });
 
-        it("variable().toString() should return SPARQL compatible syntax", function () {
+        it("variable().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual("?a", oxigraph.variable("a").toString());
         });
 
-        it("quad().toString() should return SPARQL compatible syntax", function () {
+        it("quad().toString() should return SPARQL compatible syntax", () => {
             assert.strictEqual(
                 "<http://example.com/s> <http://example.com/p> <<<http://example.com/s1> <http://example.com/p1> <http://example.com/o1>>> <http://example.com/g>",
                 oxigraph

js/test/store.mjs

@@ -1,8 +1,8 @@
 /* global describe, it */
 
-import { Store } from "../pkg/oxigraph.js";
+import assert from "node:assert";
 import dataModel from "@rdfjs/data-model";
-import assert from "assert";
+import { Store } from "../pkg/oxigraph.js";
 
 const ex = dataModel.namedNode("http://example.com");
 const triple = dataModel.quad(
@@ -11,17 +11,17 @@ const triple = dataModel.quad(
     dataModel.literal("o"),
 );
 
-describe("Store", function () {
-    describe("#add()", function () {
-        it("an added quad should be in the store", function () {
+describe("Store", () => {
+    describe("#add()", () => {
+        it("an added quad should be in the store", () => {
             const store = new Store();
             store.add(dataModel.quad(ex, ex, triple));
             assert(store.has(dataModel.quad(ex, ex, triple)));
         });
     });
 
-    describe("#delete()", function () {
-        it("an removed quad should not be in the store anymore", function () {
+    describe("#delete()", () => {
+        it("an removed quad should not be in the store anymore", () => {
             const store = new Store([dataModel.quad(triple, ex, ex)]);
             assert(store.has(dataModel.quad(triple, ex, ex)));
             store.delete(dataModel.quad(triple, ex, ex));
@@ -29,22 +29,22 @@ describe("Store", function () {
         });
     });
 
-    describe("#has()", function () {
-        it("an added quad should be in the store", function () {
+    describe("#has()", () => {
+        it("an added quad should be in the store", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             assert(store.has(dataModel.quad(ex, ex, ex)));
         });
     });
 
-    describe("#size()", function () {
-        it("A store with one quad should have 1 for size", function () {
+    describe("#size()", () => {
+        it("A store with one quad should have 1 for size", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             assert.strictEqual(1, store.size);
         });
     });
 
-    describe("#match_quads()", function () {
-        it("blank pattern should return all quads", function () {
+    describe("#match_quads()", () => {
+        it("blank pattern should return all quads", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             const results = store.match();
             assert.strictEqual(1, results.length);
@@ -52,32 +52,32 @@ describe("Store", function () {
         });
     });
 
-    describe("#query()", function () {
-        it("ASK true", function () {
+    describe("#query()", () => {
+        it("ASK true", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             assert.strictEqual(true, store.query("ASK { ?s ?s ?s }"));
         });
 
-        it("ASK false", function () {
+        it("ASK false", () => {
             const store = new Store();
             assert.strictEqual(false, store.query("ASK { FILTER(false)}"));
         });
 
-        it("CONSTRUCT", function () {
+        it("CONSTRUCT", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             const results = store.query("CONSTRUCT { ?s ?p ?o } WHERE { ?s ?p ?o }");
             assert.strictEqual(1, results.length);
             assert(dataModel.quad(ex, ex, ex).equals(results[0]));
         });
 
-        it("SELECT", function () {
+        it("SELECT", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             const results = store.query("SELECT ?s WHERE { ?s ?p ?o }");
             assert.strictEqual(1, results.length);
             assert(ex.equals(results[0].get("s")));
         });
 
-        it("SELECT with NOW()", function () {
+        it("SELECT with NOW()", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             const results = store.query(
                 "SELECT * WHERE { FILTER(2022 <= YEAR(NOW()) && YEAR(NOW()) <= 2100) }",
@@ -85,15 +85,31 @@ describe("Store", function () {
             assert.strictEqual(1, results.length);
         });
 
-        it("SELECT with RAND()", function () {
+        it("SELECT with RAND()", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             const results = store.query("SELECT (RAND() AS ?y) WHERE {}");
             assert.strictEqual(1, results.length);
         });
+
+        it("SELECT with base IRI", () => {
+            const store = new Store();
+            const results = store.query("SELECT * WHERE { BIND(<t> AS ?t) }", {
+                base_iri: "http://example.com/",
+            });
+            assert.strictEqual(1, results.length);
+        });
+
+        it("SELECT with union graph", () => {
+            const store = new Store([dataModel.quad(ex, ex, ex, ex)]);
+            const results = store.query("SELECT * WHERE { ?s ?p ?o }", {
+                use_default_graph_as_union: true,
+            });
+            assert.strictEqual(1, results.length);
+        });
     });
 
-    describe("#update()", function () {
-        it("INSERT DATA", function () {
+    describe("#update()", () => {
+        it("INSERT DATA", () => {
             const store = new Store();
             store.update(
                 "INSERT DATA { <http://example.com> <http://example.com> <http://example.com> }",
@@ -101,7 +117,7 @@ describe("Store", function () {
             assert.strictEqual(1, store.size);
         });
 
-        it("DELETE DATA", function () {
+        it("DELETE DATA", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             store.update(
                 "DELETE DATA { <http://example.com> <http://example.com> <http://example.com> }",
@@ -109,15 +125,15 @@ describe("Store", function () {
             assert.strictEqual(0, store.size);
         });
 
-        it("DELETE WHERE", function () {
+        it("DELETE WHERE", () => {
             const store = new Store([dataModel.quad(ex, ex, ex)]);
             store.update("DELETE WHERE { ?v ?v ?v }");
             assert.strictEqual(0, store.size);
         });
     });
 
-    describe("#load()", function () {
-        it("load NTriples in the default graph", function () {
+    describe("#load()", () => {
+        it("load NTriples in the default graph", () => {
             const store = new Store();
             store.load(
                 "<http://example.com> <http://example.com> <http://example.com> .",
@@ -126,7 +142,7 @@ describe("Store", function () {
             assert(store.has(dataModel.quad(ex, ex, ex)));
         });
 
-        it("load NTriples in an other graph", function () {
+        it("load NTriples in an other graph", () => {
             const store = new Store();
             store.load(
                 "<http://example.com> <http://example.com> <http://example.com> .",
@@ -137,7 +153,7 @@ describe("Store", function () {
             assert(store.has(dataModel.quad(ex, ex, ex, ex)));
         });
 
-        it("load Turtle with a base IRI", function () {
+        it("load Turtle with a base IRI", () => {
             const store = new Store();
             store.load(
                 "<http://example.com> <http://example.com> <> .",
@@ -147,7 +163,7 @@ describe("Store", function () {
             assert(store.has(dataModel.quad(ex, ex, ex)));
         });
 
-        it("load NQuads", function () {
+        it("load NQuads", () => {
             const store = new Store();
             store.load(
                 "<http://example.com> <http://example.com> <http://example.com> <http://example.com> .",
@@ -156,7 +172,7 @@ describe("Store", function () {
             assert(store.has(dataModel.quad(ex, ex, ex, ex)));
         });
 
-        it("load TriG with a base IRI", function () {
+        it("load TriG with a base IRI", () => {
             const store = new Store();
             store.load(
                 "GRAPH <> { <http://example.com> <http://example.com> <> }",
@@ -167,8 +183,8 @@ describe("Store", function () {
         });
     });
 
-    describe("#dump()", function () {
-        it("dump dataset content", function () {
+    describe("#dump()", () => {
+        it("dump dataset content", () => {
             const store = new Store([dataModel.quad(ex, ex, ex, ex)]);
             assert.strictEqual(
                 "<http://example.com> <http://example.com> <http://example.com> <http://example.com> .\n",
@@ -176,7 +192,7 @@ describe("Store", function () {
             );
         });
 
-        it("dump named graph content", function () {
+        it("dump named graph content", () => {
             const store = new Store([dataModel.quad(ex, ex, ex, ex)]);
             assert.strictEqual(
                 "<http://example.com> <http://example.com> <http://example.com> .\n",
@@ -184,9 +200,9 @@ describe("Store", function () {
             );
         });
 
-        it("dump default graph content", function () {
+        it("dump default graph content", () => {
             const store = new Store([dataModel.quad(ex, ex, ex, ex)]);
-            assert.strictEqual("", store.dump("application/n-triples"));
+            assert.strictEqual("", store.dump("application/n-triples", dataModel.defaultGraph()));
         });
     });
 });

lib/Cargo.toml

@@ -1,63 +0,0 @@
-[package]
-name = "oxigraph"
-version = "0.3.22"
-authors = ["Tpt <thomas@pellissier-tanon.fr>"]
-license = "MIT OR Apache-2.0"
-readme = "README.md"
-keywords = ["RDF", "SPARQL", "graph-database", "database"]
-categories = ["database-implementations"]
-repository = "https://github.com/oxigraph/oxigraph/tree/main/lib"
-homepage = "https://oxigraph.org/"
-documentation = "https://docs.rs/oxigraph"
-description = """
-a SPARQL database and RDF toolkit
-"""
-edition = "2021"
-rust-version = "1.60"
-
-[package.metadata.docs.rs]
-all-features = true
-
-[features]
-default = []
-http_client = ["oxhttp", "oxhttp/rustls"]
-rocksdb_debug = []
-
-[dependencies]
-rand = "0.8"
-md-5 = "0.10"
-sha-1 = "0.10"
-sha2 = "0.10"
-digest = "0.10"
-regex = "1"
-oxilangtag = "0.1"
-oxiri = "0.2"
-rio_api = "0.8"
-rio_turtle = "0.8"
-rio_xml = "0.8"
-hex = "0.4"
-siphasher = ">=0.3,<2.0"
-lazy_static = "1"
-json-event-parser = "0.1"
-oxrdf = { version = "0.1.7", path="oxrdf", features = ["rdf-star", "oxsdatatypes"] }
-oxsdatatypes = { version = "0.1.3", path="oxsdatatypes" }
-spargebra = { version = "0.2.8", path="spargebra", features = ["rdf-star", "sep-0002", "sep-0006"] }
-sparesults = { version = "0.1.8", path="sparesults", features = ["rdf-star"] }
-
-[target.'cfg(not(target_family = "wasm"))'.dependencies]
-libc = "0.2"
-oxrocksdb-sys = { version = "0.3.22", path="../oxrocksdb-sys" }
-oxhttp = { version = "0.1", optional = true }
-
-[target.'cfg(all(target_family = "wasm", target_os = "unknown"))'.dependencies]
-getrandom = { version = "0.2", features = ["js"] }
-js-sys = "0.3"
-
-[target.'cfg(not(target_family = "wasm"))'.dev-dependencies]
-criterion = "0.4"
-oxhttp = "0.1"
-zstd = "0.12"
-
-[[bench]]
-name = "store"
-harness = false

lib/README.md

@@ -1,72 +1,13 @@
-Oxigraph
-========
-
-[![Latest Version](https://img.shields.io/crates/v/oxigraph.svg)](https://crates.io/crates/oxigraph)
-[![Released API docs](https://docs.rs/oxigraph/badge.svg)](https://docs.rs/oxigraph)
-[![Crates.io downloads](https://img.shields.io/crates/d/oxigraph)](https://crates.io/crates/oxigraph)
-[![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
-[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge)
-
-Oxigraph is a graph database library implementing the [SPARQL](https://www.w3.org/TR/sparql11-overview/) standard.
-
-Its goal is to provide a compliant, safe and fast on-disk graph database.
-It also provides a set of utility functions for reading, writing, and processing RDF files.
-
-Oxigraph is in heavy development and SPARQL query evaluation has not been optimized yet.
-
-Oxigraph also provides [a standalone HTTP server](https://crates.io/crates/oxigraph_server) and [a Python library](https://pyoxigraph.readthedocs.io/) based on this library.
-
-
-Oxigraph implements the following specifications:
-* [SPARQL 1.1 Query](https://www.w3.org/TR/sparql11-query/), [SPARQL 1.1 Update](https://www.w3.org/TR/sparql11-update/), and [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/).
-* [Turtle](https://www.w3.org/TR/turtle/), [TriG](https://www.w3.org/TR/trig/), [N-Triples](https://www.w3.org/TR/n-triples/), [N-Quads](https://www.w3.org/TR/n-quads/), and [RDF XML](https://www.w3.org/TR/rdf-syntax-grammar/) RDF serialization formats for both data ingestion and retrieval using the [Rio library](https://github.com/oxigraph/rio).
-* [SPARQL Query Results XML Format](https://www.w3.org/TR/rdf-sparql-XMLres/), [SPARQL 1.1 Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/) and [SPARQL 1.1 Query Results CSV and TSV Formats](https://www.w3.org/TR/sparql11-results-csv-tsv/).
-
-A preliminary benchmark [is provided](../bench/README.md). Oxigraph internal design [is described on the wiki](https://github.com/oxigraph/oxigraph/wiki/Architecture).
-
-The main entry point of Oxigraph is the [`Store`](store::Store) struct:
-```rust
-use oxigraph::store::Store;
-use oxigraph::model::*;
-use oxigraph::sparql::QueryResults;
-
-let store = Store::new().unwrap();
-
-// insertion
-let ex = NamedNode::new("http://example.com").unwrap();
-let quad = Quad::new(ex.clone(), ex.clone(), ex.clone(), GraphName::DefaultGraph);
-store.insert(&quad).unwrap();
-
-// quad filter
-let results = store.quads_for_pattern(Some(ex.as_ref().into()), None, None, None).collect::<Result<Vec<Quad>,_>>().unwrap();
-assert_eq!(vec![quad], results);
-
-// SPARQL query
-if let QueryResults::Solutions(mut solutions) =  store.query("SELECT ?s WHERE { ?s ?p ?o }").unwrap() {
-    assert_eq!(solutions.next().unwrap().unwrap().get("s"), Some(&ex.into()));
-}
-```
-
-Some parts of this library are available as standalone crates:
-* [`oxrdf`](https://crates.io/crates/oxrdf) provides datastructures encoding RDF basic concepts (the `oxigraph::model` module).
-* [`spargebra`](https://crates.io/crates/spargebra) provides a SPARQL parser.
-* [`sparesults`](https://crates.io/crates/sparesults) provides parsers and serializers for SPARQL result formats.
-
-To build the library, don't forget to clone the submodules using `git clone --recursive https://github.com/oxigraph/oxigraph.git` to clone the repository including submodules or `git submodule update --init` to add submodules to the already cloned repository.
-
-
-## License
-
-This project is licensed under either of
-
-* Apache License, Version 2.0, ([LICENSE-APACHE](../LICENSE-APACHE) or
-  `<http://www.apache.org/licenses/LICENSE-2.0>`)
-* MIT license ([LICENSE-MIT](../LICENSE-MIT) or
-  `<http://opensource.org/licenses/MIT>`)
-
-at your option.
-
-
-### Contribution
-
-Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in Oxigraph by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.
+Oxigraph Rust crates
+====================
+
+Oxigraph is implemented in Rust.
+It is composed on a main library, [`oxigraph`](./oxigraph) and a set of smaller crates used by the `oxigraph` crate:
+* [`oxrdf`](./oxrdf), datastructures encoding RDF basic concepts (the `model` module of the `oxigraph` crate).
+* [`oxrdfio`](./oxrdfio), a unified parser and serializer API for RDF formats (the `io` module of the `oxigraph` crate). It itself relies on:
+    * [`oxttl`](./oxttl), N-Triple, N-Quad, Turtle, TriG and N3 parsing and serialization.
+    * [`oxrdfxml`](./oxrdfxml), RDF/XML parsing and serialization.
+* [`spargebra`](./spargebra), a SPARQL parser.
+* [`sparesults`](./sparesults), parsers and serializers for SPARQL result formats (the `sparql::results` module of the `oxigraph` crate).
+* [`sparopt`](./sparesults), a SPARQL optimizer.
+* [`oxsdatatypes`](./oxsdatatypes), an implementation of some XML Schema datatypes.

lib/benches/store.rs

@@ -1,265 +0,0 @@
-use criterion::{criterion_group, criterion_main, Criterion, Throughput};
-use oxhttp::model::{Method, Request, Status};
-use oxigraph::io::GraphFormat;
-use oxigraph::model::GraphNameRef;
-use oxigraph::sparql::{Query, QueryResults, Update};
-use oxigraph::store::Store;
-use rand::random;
-use std::env::temp_dir;
-use std::fs::{remove_dir_all, File};
-use std::io::{BufRead, BufReader, Cursor, Read};
-use std::path::{Path, PathBuf};
-
-fn store_load(c: &mut Criterion) {
-    {
-        let mut data = Vec::new();
-        read_data("explore-1000.nt.zst")
-            .read_to_end(&mut data)
-            .unwrap();
-
-        let mut group = c.benchmark_group("store load");
-        group.throughput(Throughput::Bytes(data.len() as u64));
-        group.sample_size(10);
-        group.bench_function("load BSBM explore 1000 in memory", |b| {
-            b.iter(|| {
-                let store = Store::new().unwrap();
-                do_load(&store, &data);
-            })
-        });
-        group.bench_function("load BSBM explore 1000 in on disk", |b| {
-            b.iter(|| {
-                let path = TempDir::default();
-                let store = Store::open(&path).unwrap();
-                do_load(&store, &data);
-            })
-        });
-        group.bench_function("load BSBM explore 1000 in on disk with bulk load", |b| {
-            b.iter(|| {
-                let path = TempDir::default();
-                let store = Store::open(&path).unwrap();
-                do_bulk_load(&store, &data);
-            })
-        });
-    }
-
-    {
-        let mut data = Vec::new();
-        read_data("explore-10000.nt.zst")
-            .read_to_end(&mut data)
-            .unwrap();
-
-        let mut group = c.benchmark_group("store load large");
-        group.throughput(Throughput::Bytes(data.len() as u64));
-        group.sample_size(10);
-        group.bench_function("load BSBM explore 10000 in on disk with bulk load", |b| {
-            b.iter(|| {
-                let path = TempDir::default();
-                let store = Store::open(&path).unwrap();
-                do_bulk_load(&store, &data);
-            })
-        });
-    }
-}
-
-fn do_load(store: &Store, data: &[u8]) {
-    store
-        .load_graph(
-            Cursor::new(&data),
-            GraphFormat::NTriples,
-            GraphNameRef::DefaultGraph,
-            None,
-        )
-        .unwrap();
-    store.optimize().unwrap();
-}
-
-fn do_bulk_load(store: &Store, data: &[u8]) {
-    store
-        .bulk_loader()
-        .load_graph(
-            Cursor::new(&data),
-            GraphFormat::NTriples,
-            GraphNameRef::DefaultGraph,
-            None,
-        )
-        .unwrap();
-    store.optimize().unwrap();
-}
-
-fn store_query_and_update(c: &mut Criterion) {
-    let mut data = Vec::new();
-    read_data("explore-1000.nt.zst")
-        .read_to_end(&mut data)
-        .unwrap();
-
-    let operations = read_data("mix-exploreAndUpdate-1000.tsv.zst")
-        .lines()
-        .map(|l| {
-            let l = l.unwrap();
-            let mut parts = l.trim().split('\t');
-            let kind = parts.next().unwrap();
-            let operation = parts.next().unwrap();
-            match kind {
-                "query" => Operation::Query(Query::parse(operation, None).unwrap()),
-                "update" => Operation::Update(Update::parse(operation, None).unwrap()),
-                _ => panic!("Unexpected operation kind {kind}"),
-            }
-        })
-        .collect::<Vec<_>>();
-    let query_operations = operations
-        .iter()
-        .filter(|o| matches!(o, Operation::Query(_)))
-        .cloned()
-        .collect::<Vec<_>>();
-
-    let mut group = c.benchmark_group("store operations");
-    group.throughput(Throughput::Elements(operations.len() as u64));
-    group.sample_size(10);
-
-    {
-        let memory_store = Store::new().unwrap();
-        do_bulk_load(&memory_store, &data);
-        group.bench_function("BSBM explore 1000 query in memory", |b| {
-            b.iter(|| run_operation(&memory_store, &query_operations))
-        });
-        group.bench_function("BSBM explore 1000 queryAndUpdate in memory", |b| {
-            b.iter(|| run_operation(&memory_store, &operations))
-        });
-    }
-
-    {
-        let path = TempDir::default();
-        let disk_store = Store::open(&path).unwrap();
-        do_bulk_load(&disk_store, &data);
-        group.bench_function("BSBM explore 1000 query on disk", |b| {
-            b.iter(|| run_operation(&disk_store, &query_operations))
-        });
-        group.bench_function("BSBM explore 1000 queryAndUpdate on disk", |b| {
-            b.iter(|| run_operation(&disk_store, &operations))
-        });
-    }
-}
-
-fn run_operation(store: &Store, operations: &[Operation]) {
-    for operation in operations {
-        match operation {
-            Operation::Query(q) => match store.query(q.clone()).unwrap() {
-                QueryResults::Boolean(_) => (),
-                QueryResults::Solutions(s) => {
-                    for s in s {
-                        s.unwrap();
-                    }
-                }
-                QueryResults::Graph(g) => {
-                    for t in g {
-                        t.unwrap();
-                    }
-                }
-            },
-            Operation::Update(u) => store.update(u.clone()).unwrap(),
-        }
-    }
-}
-
-fn sparql_parsing(c: &mut Criterion) {
-    let mut data = Vec::new();
-    read_data("explore-1000.nt.zst")
-        .read_to_end(&mut data)
-        .unwrap();
-
-    let operations = read_data("mix-exploreAndUpdate-1000.tsv.zst")
-        .lines()
-        .map(|l| {
-            let l = l.unwrap();
-            let mut parts = l.trim().split('\t');
-            let kind = parts.next().unwrap();
-            let operation = parts.next().unwrap();
-            match kind {
-                "query" => RawOperation::Query(operation.to_owned()),
-                "update" => RawOperation::Update(operation.to_owned()),
-                _ => panic!("Unexpected operation kind {kind}"),
-            }
-        })
-        .collect::<Vec<_>>();
-
-    let mut group = c.benchmark_group("sparql parsing");
-    group.sample_size(10);
-    group.throughput(Throughput::Bytes(
-        operations
-            .iter()
-            .map(|o| match o {
-                RawOperation::Query(q) => q.len(),
-                RawOperation::Update(u) => u.len(),
-            })
-            .sum::<usize>() as u64,
-    ));
-    group.bench_function("BSBM query and update set", |b| {
-        b.iter(|| {
-            for operation in &operations {
-                match operation {
-                    RawOperation::Query(q) => {
-                        Query::parse(q, None).unwrap();
-                    }
-                    RawOperation::Update(u) => {
-                        Update::parse(u, None).unwrap();
-                    }
-                }
-            }
-        })
-    });
-}
-
-criterion_group!(store, sparql_parsing, store_query_and_update, store_load);
-
-criterion_main!(store);
-
-fn read_data(file: &str) -> impl BufRead {
-    if !Path::new(file).exists() {
-        let mut client = oxhttp::Client::new();
-        client.set_redirection_limit(5);
-        let url = format!("https://github.com/Tpt/bsbm-tools/releases/download/v0.2/{file}");
-        let request = Request::builder(Method::GET, url.parse().unwrap()).build();
-        let response = client.request(request).unwrap();
-        assert_eq!(
-            response.status(),
-            Status::OK,
-            "{}",
-            response.into_body().to_string().unwrap()
-        );
-        std::io::copy(&mut response.into_body(), &mut File::create(file).unwrap()).unwrap();
-    }
-    BufReader::new(zstd::Decoder::new(File::open(file).unwrap()).unwrap())
-}
-
-#[derive(Clone)]
-enum RawOperation {
-    Query(String),
-    Update(String),
-}
-
-#[allow(clippy::large_enum_variant)]
-#[derive(Clone)]
-enum Operation {
-    Query(Query),
-    Update(Update),
-}
-
-struct TempDir(PathBuf);
-
-impl Default for TempDir {
-    fn default() -> Self {
-        Self(temp_dir().join(format!("oxigraph-bench-{}", random::<u128>())))
-    }
-}
-
-impl AsRef<Path> for TempDir {
-    fn as_ref(&self) -> &Path {
-        &self.0
-    }
-}
-
-impl Drop for TempDir {
-    fn drop(&mut self) {
-        remove_dir_all(&self.0).unwrap()
-    }
-}

lib/oxigraph/Cargo.toml

@@ -0,0 +1,59 @@
+[package]
+name = "oxigraph"
+version.workspace = true
+authors.workspace = true
+license.workspace = true
+readme = "README.md"
+keywords = ["RDF", "SPARQL", "graph-database", "database"]
+categories = ["database-implementations"]
+repository = "https://github.com/oxigraph/oxigraph/tree/main/lib/oxigraph"
+homepage = "https://oxigraph.org/"
+documentation = "https://docs.rs/oxigraph"
+description = """
+a SPARQL database and RDF toolkit
+"""
+edition.workspace = true
+rust-version.workspace = true
+
+[features]
+js = ["getrandom/js", "oxsdatatypes/js", "js-sys"]
+
+
+[dependencies]
+digest.workspace = true
+hex.workspace = true
+json-event-parser.workspace = true
+md-5.workspace = true
+oxilangtag.workspace = true
+oxiri.workspace = true
+oxrdf = { workspace = true, features = ["rdf-star", "oxsdatatypes"] }
+oxrdfio = { workspace = true, features = ["rdf-star"] }
+oxsdatatypes.workspace = true
+rand.workspace = true
+regex.workspace = true
+sha1.workspace = true
+sha2.workspace = true
+siphasher.workspace = true
+sparesults = { workspace = true, features = ["rdf-star"] }
+spargebra = { workspace = true, features = ["rdf-star", "sep-0002", "sep-0006"] }
+sparopt = { workspace = true, features = ["rdf-star", "sep-0002", "sep-0006"] }
+thiserror.workspace = true
+
+[target.'cfg(not(target_family = "wasm"))'.dependencies]
+libc = "0.2"
+rocksdb.workspace = true
+
+[target.'cfg(all(target_family = "wasm", target_os = "unknown"))'.dependencies]
+getrandom.workspace = true
+js-sys = { workspace = true, optional = true }
+
+[target.'cfg(not(target_family = "wasm"))'.dev-dependencies]
+codspeed-criterion-compat.workspace = true
+zstd.workspace = true
+
+[lints]
+workspace = true
+
+[package.metadata.docs.rs]
+rustdoc-args = ["--cfg", "docsrs"]
+

lib/oxigraph/README.md

@@ -0,0 +1,82 @@
+Oxigraph
+========
+
+[![Latest Version](https://img.shields.io/crates/v/oxigraph.svg)](https://crates.io/crates/oxigraph)
+[![Released API docs](https://docs.rs/oxigraph/badge.svg)](https://docs.rs/oxigraph)
+[![Crates.io downloads](https://img.shields.io/crates/d/oxigraph)](https://crates.io/crates/oxigraph)
+[![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
+[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community)
+
+Oxigraph is a graph database library implementing the [SPARQL](https://www.w3.org/TR/sparql11-overview/) standard.
+
+Its goal is to provide a compliant, safe and fast on-disk graph database.
+It also provides a set of utility functions for reading, writing, and processing RDF files.
+
+Oxigraph is in heavy development and SPARQL query evaluation has not been optimized yet.
+
+Oxigraph also provides [a CLI tool](https://crates.io/crates/oxigraph-cli) and [a Python library](https://pyoxigraph.readthedocs.io/) based on this library.
+
+
+Oxigraph implements the following specifications:
+* [SPARQL 1.1 Query](https://www.w3.org/TR/sparql11-query/), [SPARQL 1.1 Update](https://www.w3.org/TR/sparql11-update/), and [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/).
+* [Turtle](https://www.w3.org/TR/turtle/), [TriG](https://www.w3.org/TR/trig/), [N-Triples](https://www.w3.org/TR/n-triples/), [N-Quads](https://www.w3.org/TR/n-quads/), and [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/) RDF serialization formats for both data ingestion and retrieval.
+* [SPARQL Query Results XML Format](https://www.w3.org/TR/rdf-sparql-XMLres/), [SPARQL 1.1 Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/) and [SPARQL 1.1 Query Results CSV and TSV Formats](https://www.w3.org/TR/sparql11-results-csv-tsv/).
+
+A preliminary benchmark [is provided](../bench/README.md). Oxigraph internal design [is described on the wiki](https://github.com/oxigraph/oxigraph/wiki/Architecture).
+
+The main entry point of Oxigraph is the [`Store`](store::Store) struct:
+```rust
+use oxigraph::store::Store;
+use oxigraph::model::*;
+use oxigraph::sparql::QueryResults;
+
+let store = Store::new().unwrap();
+
+// insertion
+let ex = NamedNode::new("http://example.com").unwrap();
+let quad = Quad::new(ex.clone(), ex.clone(), ex.clone(), GraphName::DefaultGraph);
+store.insert(&quad).unwrap();
+
+// quad filter
+let results = store.quads_for_pattern(Some(ex.as_ref().into()), None, None, None).collect::<Result<Vec<Quad>,_>>().unwrap();
+assert_eq!(vec![quad], results);
+
+// SPARQL query
+if let QueryResults::Solutions(mut solutions) =  store.query("SELECT ?s WHERE { ?s ?p ?o }").unwrap() {
+    assert_eq!(solutions.next().unwrap().unwrap().get("s"), Some(&ex.into()));
+}
+```
+
+It is based on these crates that can be used separately:
+* [`oxrdf`](https://crates.io/crates/oxrdf), datastructures encoding RDF basic concepts (the [`oxigraph::model`](crate::model) module).
+* [`oxrdfio`](https://crates.io/crates/oxrdfio), a unified parser and serializer API for RDF formats (the [`oxigraph::io`](crate::io) module). It itself relies on:
+  * [`oxttl`](https://crates.io/crates/oxttl), N-Triple, N-Quad, Turtle, TriG and N3 parsing and serialization.
+  * [`oxrdfxml`](https://crates.io/crates/oxrdfxml), RDF/XML parsing and serialization.
+* [`spargebra`](https://crates.io/crates/spargebra), a SPARQL parser.
+* [`sparesults`](https://crates.io/crates/sparesults), parsers and serializers for SPARQL result formats (the [`oxigraph::sparql::results`](crate::sparql::results) module).
+* [`sparopt`](https://crates.io/crates/sparesults), a SPARQL optimizer.
+* [`oxsdatatypes`](https://crates.io/crates/oxsdatatypes), an implementation of some XML Schema datatypes.
+
+To build the library locally, don't forget to clone the submodules using `git clone --recursive https://github.com/oxigraph/oxigraph.git` to clone the repository including submodules or `git submodule update --init` to add submodules to the already cloned repository.
+
+It is possible to disable the RocksDB storage backend to only use the in-memory fallback by disabling the `rocksdb` default feature:
+```toml
+oxigraph = { version = "*", default-features = false }
+```
+This is the default behavior when compiling Oxigraph to WASM.
+
+## License
+
+This project is licensed under either of
+
+* Apache License, Version 2.0, ([LICENSE-APACHE](../LICENSE-APACHE) or
+  `<http://www.apache.org/licenses/LICENSE-2.0>`)
+* MIT license ([LICENSE-MIT](../LICENSE-MIT) or
+  `<http://opensource.org/licenses/MIT>`)
+
+at your option.
+
+
+### Contribution
+
+Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in Oxigraph by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

lib/oxigraph/src/io/format.rs

@@ -1,8 +1,13 @@
+#![allow(deprecated)]
+
+use oxrdfio::{RdfFormat, RdfParser, RdfSerializer};
+
 /// [RDF graph](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-graph) serialization formats.
 ///
 /// This enumeration is non exhaustive. New formats like JSON-LD will be added in the future.
 #[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
 #[non_exhaustive]
+#[deprecated(note = "use RdfFormat instead", since = "0.4.0")]
 pub enum GraphFormat {
     /// [N-Triples](https://www.w3.org/TR/n-triples/)
     NTriples,
@@ -18,7 +23,10 @@ impl GraphFormat {
     /// ```
     /// use oxigraph::io::GraphFormat;
     ///
-    /// assert_eq!(GraphFormat::NTriples.iri(), "http://www.w3.org/ns/formats/N-Triples")
+    /// assert_eq!(
+    ///     GraphFormat::NTriples.iri(),
+    ///     "http://www.w3.org/ns/formats/N-Triples"
+    /// )
     /// ```
     #[inline]
     pub fn iri(self) -> &'static str {
@@ -60,6 +68,7 @@ impl GraphFormat {
             Self::RdfXml => "rdf",
         }
     }
+
     /// Looks for a known format from a media type.
     ///
     /// It supports some media type aliases.
@@ -69,7 +78,10 @@ impl GraphFormat {
     /// ```
     /// use oxigraph::io::GraphFormat;
     ///
-    /// assert_eq!(GraphFormat::from_media_type("text/turtle; charset=utf-8"), Some(GraphFormat::Turtle))
+    /// assert_eq!(
+    ///     GraphFormat::from_media_type("text/turtle; charset=utf-8"),
+    ///     Some(GraphFormat::Turtle)
+    /// )
     /// ```
     #[inline]
     pub fn from_media_type(media_type: &str) -> Option<Self> {
@@ -89,7 +101,10 @@ impl GraphFormat {
     /// ```
     /// use oxigraph::io::GraphFormat;
     ///
-    /// assert_eq!(GraphFormat::from_extension("nt"), Some(GraphFormat::NTriples))
+    /// assert_eq!(
+    ///     GraphFormat::from_extension("nt"),
+    ///     Some(GraphFormat::NTriples)
+    /// )
     /// ```
     #[inline]
     pub fn from_extension(extension: &str) -> Option<Self> {
@@ -102,11 +117,37 @@ impl GraphFormat {
     }
 }
 
+impl From<GraphFormat> for RdfFormat {
+    #[inline]
+    fn from(format: GraphFormat) -> Self {
+        match format {
+            GraphFormat::NTriples => Self::NTriples,
+            GraphFormat::Turtle => Self::Turtle,
+            GraphFormat::RdfXml => Self::RdfXml,
+        }
+    }
+}
+
+impl From<GraphFormat> for RdfParser {
+    #[inline]
+    fn from(format: GraphFormat) -> Self {
+        RdfFormat::from(format).into()
+    }
+}
+
+impl From<GraphFormat> for RdfSerializer {
+    #[inline]
+    fn from(format: GraphFormat) -> Self {
+        RdfFormat::from(format).into()
+    }
+}
+
 /// [RDF dataset](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-dataset) serialization formats.
 ///
 /// This enumeration is non exhaustive. New formats like JSON-LD will be added in the future.
 #[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
 #[non_exhaustive]
+#[deprecated(note = "use RdfFormat instead", since = "0.4.0")]
 pub enum DatasetFormat {
     /// [N-Quads](https://www.w3.org/TR/n-quads/)
     NQuads,
@@ -120,7 +161,10 @@ impl DatasetFormat {
     /// ```
     /// use oxigraph::io::DatasetFormat;
     ///
-    /// assert_eq!(DatasetFormat::NQuads.iri(), "http://www.w3.org/ns/formats/N-Quads")
+    /// assert_eq!(
+    ///     DatasetFormat::NQuads.iri(),
+    ///     "http://www.w3.org/ns/formats/N-Quads"
+    /// )
     /// ```
     #[inline]
     pub fn iri(self) -> &'static str {
@@ -159,6 +203,7 @@ impl DatasetFormat {
             Self::TriG => "trig",
         }
     }
+
     /// Looks for a known format from a media type.
     ///
     /// It supports some media type aliases.
@@ -167,7 +212,10 @@ impl DatasetFormat {
     /// ```
     /// use oxigraph::io::DatasetFormat;
     ///
-    /// assert_eq!(DatasetFormat::from_media_type("application/n-quads; charset=utf-8"), Some(DatasetFormat::NQuads))
+    /// assert_eq!(
+    ///     DatasetFormat::from_media_type("application/n-quads; charset=utf-8"),
+    ///     Some(DatasetFormat::NQuads)
+    /// )
     /// ```
     #[inline]
     pub fn from_media_type(media_type: &str) -> Option<Self> {
@@ -186,7 +234,10 @@ impl DatasetFormat {
     /// ```
     /// use oxigraph::io::DatasetFormat;
     ///
-    /// assert_eq!(DatasetFormat::from_extension("nq"), Some(DatasetFormat::NQuads))
+    /// assert_eq!(
+    ///     DatasetFormat::from_extension("nq"),
+    ///     Some(DatasetFormat::NQuads)
+    /// )
     /// ```
     #[inline]
     pub fn from_extension(extension: &str) -> Option<Self> {
@@ -198,12 +249,36 @@ impl DatasetFormat {
     }
 }
 
+impl From<DatasetFormat> for RdfFormat {
+    #[inline]
+    fn from(format: DatasetFormat) -> Self {
+        match format {
+            DatasetFormat::NQuads => Self::NQuads,
+            DatasetFormat::TriG => Self::TriG,
+        }
+    }
+}
+
+impl From<DatasetFormat> for RdfParser {
+    #[inline]
+    fn from(format: DatasetFormat) -> Self {
+        RdfFormat::from(format).into()
+    }
+}
+
+impl From<DatasetFormat> for RdfSerializer {
+    #[inline]
+    fn from(format: DatasetFormat) -> Self {
+        RdfFormat::from(format).into()
+    }
+}
+
 impl TryFrom<DatasetFormat> for GraphFormat {
     type Error = ();
 
     /// Attempts to find a graph format that is a subset of this [`DatasetFormat`].
     #[inline]
-    fn try_from(value: DatasetFormat) -> Result<Self, ()> {
+    fn try_from(value: DatasetFormat) -> Result<Self, Self::Error> {
         match value {
             DatasetFormat::NQuads => Ok(Self::NTriples),
             DatasetFormat::TriG => Ok(Self::Turtle),
@@ -216,7 +291,7 @@ impl TryFrom<GraphFormat> for DatasetFormat {
 
     /// Attempts to find a dataset format that is a superset of this [`GraphFormat`].
     #[inline]
-    fn try_from(value: GraphFormat) -> Result<Self, ()> {
+    fn try_from(value: GraphFormat) -> Result<Self, Self::Error> {
         match value {
             GraphFormat::NTriples => Ok(Self::NQuads),
             GraphFormat::Turtle => Ok(Self::TriG),

lib/oxigraph/src/io/mod.rs

@@ -0,0 +1,39 @@
+//! Utilities to read and write RDF graphs and datasets using [OxRDF I/O](https://crates.io/crates/oxrdfio).
+//!
+//! The entry points of this module are the two [`RdfParser`] and [`RdfSerializer`] structs.
+//!
+//! Usage example converting a Turtle file to a N-Triples file:
+//! ```
+//! use oxigraph::io::{RdfFormat, RdfParser, RdfSerializer};
+//!
+//! let turtle_file = b"@base <http://example.com/> .
+//! @prefix schema: <http://schema.org/> .
+//! <foo> a schema:Person ;
+//!     schema:name \"Foo\" .
+//! <bar> a schema:Person ;
+//!     schema:name \"Bar\" .";
+//!
+//! let ntriples_file = b"<http://example.com/foo> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Person> .
+//! <http://example.com/foo> <http://schema.org/name> \"Foo\" .
+//! <http://example.com/bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Person> .
+//! <http://example.com/bar> <http://schema.org/name> \"Bar\" .
+//! ";
+//!
+//! let mut writer = RdfSerializer::from_format(RdfFormat::NTriples).serialize_to_write(Vec::new());
+//! for quad in RdfParser::from_format(RdfFormat::Turtle).parse_read(turtle_file.as_ref()) {
+//!     writer.write_quad(&quad.unwrap()).unwrap();
+//! }
+//! assert_eq!(writer.finish().unwrap(), ntriples_file);
+//! ```
+
+mod format;
+pub mod read;
+pub mod write;
+
+#[allow(deprecated)]
+pub use self::format::{DatasetFormat, GraphFormat};
+#[allow(deprecated)]
+pub use self::read::{DatasetParser, GraphParser};
+#[allow(deprecated)]
+pub use self::write::{DatasetSerializer, GraphSerializer};
+pub use oxrdfio::*;

lib/oxigraph/src/io/read.rs

@@ -0,0 +1,199 @@
+#![allow(deprecated)]
+
+//! Utilities to read RDF graphs and datasets.
+
+use crate::io::{DatasetFormat, GraphFormat};
+use crate::model::*;
+use oxrdfio::{FromReadQuadReader, RdfParseError, RdfParser};
+use std::io::Read;
+
+/// Parsers for RDF graph serialization formats.
+///
+/// It currently supports the following formats:
+/// * [N-Triples](https://www.w3.org/TR/n-triples/) ([`GraphFormat::NTriples`])
+/// * [Turtle](https://www.w3.org/TR/turtle/) ([`GraphFormat::Turtle`])
+/// * [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/) ([`GraphFormat::RdfXml`])
+///
+/// ```
+/// use oxigraph::io::{GraphFormat, GraphParser};
+///
+/// let file = "<http://example.com/s> <http://example.com/p> <http://example.com/o> .";
+///
+/// let parser = GraphParser::from_format(GraphFormat::NTriples);
+/// let triples = parser
+///     .read_triples(file.as_bytes())
+///     .collect::<Result<Vec<_>, _>>()?;
+///
+/// assert_eq!(triples.len(), 1);
+/// assert_eq!(triples[0].subject.to_string(), "<http://example.com/s>");
+/// # std::io::Result::Ok(())
+/// ```
+#[deprecated(note = "use RdfParser instead", since = "0.4.0")]
+pub struct GraphParser {
+    inner: RdfParser,
+}
+
+impl GraphParser {
+    /// Builds a parser for the given format.
+    #[inline]
+    pub fn from_format(format: GraphFormat) -> Self {
+        Self {
+            inner: RdfParser::from_format(format.into())
+                .without_named_graphs()
+                .rename_blank_nodes(),
+        }
+    }
+
+    /// Provides an IRI that could be used to resolve the file relative IRIs.
+    ///
+    /// ```
+    /// use oxigraph::io::{GraphFormat, GraphParser};
+    ///
+    /// let file = "</s> </p> </o> .";
+    ///
+    /// let parser =
+    ///     GraphParser::from_format(GraphFormat::Turtle).with_base_iri("http://example.com")?;
+    /// let triples = parser
+    ///     .read_triples(file.as_bytes())
+    ///     .collect::<Result<Vec<_>, _>>()?;
+    ///
+    /// assert_eq!(triples.len(), 1);
+    /// assert_eq!(triples[0].subject.to_string(), "<http://example.com/s>");
+    /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+    /// ```
+    #[inline]
+    pub fn with_base_iri(self, base_iri: impl Into<String>) -> Result<Self, IriParseError> {
+        Ok(Self {
+            inner: self.inner.with_base_iri(base_iri)?,
+        })
+    }
+
+    /// Executes the parsing itself on a [`Read`] implementation and returns an iterator of triples.
+    pub fn read_triples<R: Read>(self, reader: R) -> TripleReader<R> {
+        TripleReader {
+            parser: self.inner.parse_read(reader),
+        }
+    }
+}
+
+/// An iterator yielding read triples.
+/// Could be built using a [`GraphParser`].
+///
+/// ```
+/// use oxigraph::io::{GraphFormat, GraphParser};
+///
+/// let file = "<http://example.com/s> <http://example.com/p> <http://example.com/o> .";
+///
+/// let parser = GraphParser::from_format(GraphFormat::NTriples);
+/// let triples = parser
+///     .read_triples(file.as_bytes())
+///     .collect::<Result<Vec<_>, _>>()?;
+///
+/// assert_eq!(triples.len(), 1);
+/// assert_eq!(triples[0].subject.to_string(), "<http://example.com/s>");
+/// # std::io::Result::Ok(())
+/// ```
+#[must_use]
+pub struct TripleReader<R: Read> {
+    parser: FromReadQuadReader<R>,
+}
+
+impl<R: Read> Iterator for TripleReader<R> {
+    type Item = Result<Triple, RdfParseError>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        Some(self.parser.next()?.map(Into::into).map_err(Into::into))
+    }
+}
+
+/// A parser for RDF dataset serialization formats.
+///
+/// It currently supports the following formats:
+/// * [N-Quads](https://www.w3.org/TR/n-quads/) ([`DatasetFormat::NQuads`])
+/// * [TriG](https://www.w3.org/TR/trig/) ([`DatasetFormat::TriG`])
+///
+/// ```
+/// use oxigraph::io::{DatasetFormat, DatasetParser};
+///
+/// let file = "<http://example.com/s> <http://example.com/p> <http://example.com/o> <http://example.com/g> .";
+///
+/// let parser = DatasetParser::from_format(DatasetFormat::NQuads);
+/// let quads = parser.read_quads(file.as_bytes()).collect::<Result<Vec<_>,_>>()?;
+///
+/// assert_eq!(quads.len(), 1);
+/// assert_eq!(quads[0].subject.to_string(), "<http://example.com/s>");
+/// # std::io::Result::Ok(())
+/// ```
+#[deprecated(note = "use RdfParser instead", since = "0.4.0")]
+pub struct DatasetParser {
+    inner: RdfParser,
+}
+
+impl DatasetParser {
+    /// Builds a parser for the given format.
+    #[inline]
+    pub fn from_format(format: DatasetFormat) -> Self {
+        Self {
+            inner: RdfParser::from_format(format.into()).rename_blank_nodes(),
+        }
+    }
+
+    /// Provides an IRI that could be used to resolve the file relative IRIs.
+    ///
+    /// ```
+    /// use oxigraph::io::{DatasetFormat, DatasetParser};
+    ///
+    /// let file = "<g> { </s> </p> </o> }";
+    ///
+    /// let parser =
+    ///     DatasetParser::from_format(DatasetFormat::TriG).with_base_iri("http://example.com")?;
+    /// let triples = parser
+    ///     .read_quads(file.as_bytes())
+    ///     .collect::<Result<Vec<_>, _>>()?;
+    ///
+    /// assert_eq!(triples.len(), 1);
+    /// assert_eq!(triples[0].subject.to_string(), "<http://example.com/s>");
+    /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+    /// ```
+    #[inline]
+    pub fn with_base_iri(self, base_iri: impl Into<String>) -> Result<Self, IriParseError> {
+        Ok(Self {
+            inner: self.inner.with_base_iri(base_iri)?,
+        })
+    }
+
+    /// Executes the parsing itself on a [`Read`] implementation and returns an iterator of quads.
+    pub fn read_quads<R: Read>(self, reader: R) -> QuadReader<R> {
+        QuadReader {
+            parser: self.inner.parse_read(reader),
+        }
+    }
+}
+
+/// An iterator yielding read quads.
+/// Could be built using a [`DatasetParser`].
+///
+/// ```
+/// use oxigraph::io::{DatasetFormat, DatasetParser};
+///
+/// let file = "<http://example.com/s> <http://example.com/p> <http://example.com/o> <http://example.com/g> .";
+///
+/// let parser = DatasetParser::from_format(DatasetFormat::NQuads);
+/// let quads = parser.read_quads(file.as_bytes()).collect::<Result<Vec<_>,_>>()?;
+///
+/// assert_eq!(quads.len(), 1);
+/// assert_eq!(quads[0].subject.to_string(), "<http://example.com/s>");
+/// # std::io::Result::Ok(())
+/// ```
+#[must_use]
+pub struct QuadReader<R: Read> {
+    parser: FromReadQuadReader<R>,
+}
+
+impl<R: Read> Iterator for QuadReader<R> {
+    type Item = Result<Quad, RdfParseError>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        Some(self.parser.next()?.map_err(Into::into))
+    }
+}

lib/oxigraph/src/io/write.rs

@@ -0,0 +1,185 @@
+#![allow(deprecated)]
+
+//! Utilities to write RDF graphs and datasets.
+
+use crate::io::{DatasetFormat, GraphFormat};
+use crate::model::*;
+use oxrdfio::{RdfSerializer, ToWriteQuadWriter};
+use std::io::{self, Write};
+
+/// A serializer for RDF graph serialization formats.
+///
+/// It currently supports the following formats:
+/// * [N-Triples](https://www.w3.org/TR/n-triples/) ([`GraphFormat::NTriples`])
+/// * [Turtle](https://www.w3.org/TR/turtle/) ([`GraphFormat::Turtle`])
+/// * [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/) ([`GraphFormat::RdfXml`])
+///
+/// ```
+/// use oxigraph::io::{GraphFormat, GraphSerializer};
+/// use oxigraph::model::*;
+///
+/// let mut buffer = Vec::new();
+/// let mut writer = GraphSerializer::from_format(GraphFormat::NTriples).triple_writer(&mut buffer);
+/// writer.write(&Triple {
+///     subject: NamedNode::new("http://example.com/s")?.into(),
+///     predicate: NamedNode::new("http://example.com/p")?,
+///     object: NamedNode::new("http://example.com/o")?.into(),
+/// })?;
+/// writer.finish()?;
+///
+/// assert_eq!(
+///     buffer.as_slice(),
+///     "<http://example.com/s> <http://example.com/p> <http://example.com/o> .\n".as_bytes()
+/// );
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+#[deprecated(note = "use RdfSerializer instead", since = "0.4.0")]
+pub struct GraphSerializer {
+    inner: RdfSerializer,
+}
+
+impl GraphSerializer {
+    /// Builds a serializer for the given format
+    #[inline]
+    pub fn from_format(format: GraphFormat) -> Self {
+        Self {
+            inner: RdfSerializer::from_format(format.into()),
+        }
+    }
+
+    /// Returns a [`TripleWriter`] allowing writing triples into the given [`Write`] implementation
+    pub fn triple_writer<W: Write>(self, write: W) -> TripleWriter<W> {
+        TripleWriter {
+            writer: self.inner.serialize_to_write(write),
+        }
+    }
+}
+
+/// Allows writing triples.
+/// Could be built using a [`GraphSerializer`].
+///
+/// <div class="warning">
+///
+/// Do not forget to run the [`finish`](TripleWriter::finish()) method to properly write the last bytes of the file.</div>
+///
+/// ```
+/// use oxigraph::io::{GraphFormat, GraphSerializer};
+/// use oxigraph::model::*;
+///
+/// let mut buffer = Vec::new();
+/// let mut writer = GraphSerializer::from_format(GraphFormat::NTriples).triple_writer(&mut buffer);
+/// writer.write(&Triple {
+///     subject: NamedNode::new("http://example.com/s")?.into(),
+///     predicate: NamedNode::new("http://example.com/p")?,
+///     object: NamedNode::new("http://example.com/o")?.into(),
+/// })?;
+/// writer.finish()?;
+///
+/// assert_eq!(
+///     buffer.as_slice(),
+///     "<http://example.com/s> <http://example.com/p> <http://example.com/o> .\n".as_bytes()
+/// );
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+#[must_use]
+pub struct TripleWriter<W: Write> {
+    writer: ToWriteQuadWriter<W>,
+}
+
+impl<W: Write> TripleWriter<W> {
+    /// Writes a triple
+    pub fn write<'a>(&mut self, triple: impl Into<TripleRef<'a>>) -> io::Result<()> {
+        self.writer.write_triple(triple)
+    }
+
+    /// Writes the last bytes of the file
+    pub fn finish(self) -> io::Result<()> {
+        self.writer.finish()?.flush()
+    }
+}
+
+/// A serializer for RDF graph serialization formats.
+///
+/// It currently supports the following formats:
+/// * [N-Quads](https://www.w3.org/TR/n-quads/) ([`DatasetFormat::NQuads`])
+/// * [TriG](https://www.w3.org/TR/trig/) ([`DatasetFormat::TriG`])
+///
+/// ```
+/// use oxigraph::io::{DatasetFormat, DatasetSerializer};
+/// use oxigraph::model::*;
+///
+/// let mut buffer = Vec::new();
+/// let mut writer = DatasetSerializer::from_format(DatasetFormat::NQuads).quad_writer(&mut buffer);
+/// writer.write(&Quad {
+///    subject: NamedNode::new("http://example.com/s")?.into(),
+///    predicate: NamedNode::new("http://example.com/p")?,
+///    object: NamedNode::new("http://example.com/o")?.into(),
+///    graph_name: NamedNode::new("http://example.com/g")?.into(),
+/// })?;
+/// writer.finish()?;
+///
+/// assert_eq!(buffer.as_slice(), "<http://example.com/s> <http://example.com/p> <http://example.com/o> <http://example.com/g> .\n".as_bytes());
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+#[deprecated(note = "use RdfSerializer instead", since = "0.4.0")]
+pub struct DatasetSerializer {
+    inner: RdfSerializer,
+}
+
+impl DatasetSerializer {
+    /// Builds a serializer for the given format
+    #[inline]
+    pub fn from_format(format: DatasetFormat) -> Self {
+        Self {
+            inner: RdfSerializer::from_format(format.into()),
+        }
+    }
+
+    /// Returns a [`QuadWriter`] allowing writing triples into the given [`Write`] implementation
+    pub fn quad_writer<W: Write>(self, write: W) -> QuadWriter<W> {
+        QuadWriter {
+            writer: self.inner.serialize_to_write(write),
+        }
+    }
+}
+
+/// Allows writing triples.
+/// Could be built using a [`DatasetSerializer`].
+///
+/// <div class="warning">
+///
+/// Do not forget to run the [`finish`](QuadWriter::finish()) method to properly write the last bytes of the file.</div>
+///
+/// ```
+/// use oxigraph::io::{DatasetFormat, DatasetSerializer};
+/// use oxigraph::model::*;
+///
+/// let mut buffer = Vec::new();
+/// let mut writer = DatasetSerializer::from_format(DatasetFormat::NQuads).quad_writer(&mut buffer);
+/// writer.write(&Quad {
+///    subject: NamedNode::new("http://example.com/s")?.into(),
+///    predicate: NamedNode::new("http://example.com/p")?,
+///    object: NamedNode::new("http://example.com/o")?.into(),
+///    graph_name: NamedNode::new("http://example.com/g")?.into(),
+/// })?;
+/// writer.finish()?;
+///
+/// assert_eq!(buffer.as_slice(), "<http://example.com/s> <http://example.com/p> <http://example.com/o> <http://example.com/g> .\n".as_bytes());
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+#[must_use]
+pub struct QuadWriter<W: Write> {
+    writer: ToWriteQuadWriter<W>,
+}
+
+impl<W: Write> QuadWriter<W> {
+    /// Writes a quad
+    pub fn write<'a>(&mut self, quad: impl Into<QuadRef<'a>>) -> io::Result<()> {
+        self.writer.write_quad(quad)
+    }
+
+    /// Writes the last bytes of the file
+    pub fn finish(self) -> io::Result<()> {
+        self.writer.finish()?.flush()
+    }
+}

lib/oxigraph/src/lib.rs

@@ -0,0 +1,12 @@
+#![doc = include_str!("../README.md")]
+#![doc(test(attr(deny(warnings))))]
+#![doc(test(attr(allow(deprecated))))]
+#![cfg_attr(docsrs, feature(doc_auto_cfg))]
+#![doc(html_favicon_url = "https://raw.githubusercontent.com/oxigraph/oxigraph/main/logo.svg")]
+#![doc(html_logo_url = "https://raw.githubusercontent.com/oxigraph/oxigraph/main/logo.svg")]
+
+pub mod io;
+pub mod model;
+pub mod sparql;
+mod storage;
+pub mod store;

lib/oxigraph/src/model.rs

@@ -0,0 +1,22 @@
+//! Implements data structures for [RDF 1.1 Concepts](https://www.w3.org/TR/rdf11-concepts/) using [OxRDF](https://crates.io/crates/oxrdf).
+//!
+//! Usage example:
+//!
+//! ```
+//! use oxigraph::model::*;
+//!
+//! let mut graph = Graph::default();
+//!
+//! // insertion
+//! let ex = NamedNodeRef::new("http://example.com").unwrap();
+//! let triple = TripleRef::new(ex, ex, ex);
+//! graph.insert(triple);
+//!
+//! // simple filter
+//! let results: Vec<_> = graph.triples_for_subject(ex).collect();
+//! assert_eq!(vec![triple], results);
+//! ```
+
+pub use oxrdf::*;
+
+pub use spargebra::term::GroundQuad;

lib/oxigraph/src/sparql/algebra.rs

@@ -1,15 +1,13 @@
 //! [SPARQL 1.1 Query Algebra](https://www.w3.org/TR/sparql11-query/#sparqlQuery)
 //!
 //! The root type for SPARQL queries is [`Query`] and the root type for updates is [`Update`].
-//!
-//! Warning: this implementation is an unstable work in progress
 
 use crate::model::*;
 use crate::sparql::eval::Timer;
+use oxsdatatypes::DayTimeDuration;
 use spargebra::GraphUpdateOperation;
 use std::fmt;
 use std::str::FromStr;
-use std::time::Duration;
 
 /// A parsed [SPARQL query](https://www.w3.org/TR/sparql11-query/).
 ///
@@ -25,25 +23,31 @@ use std::time::Duration;
 /// // We edit the query dataset specification
 /// let default = vec![NamedNode::new("http://example.com")?.into()];
 /// query.dataset_mut().set_default_graph(default.clone());
-/// assert_eq!(query.dataset().default_graph_graphs(), Some(default.as_slice()));
+/// assert_eq!(
+///     query.dataset().default_graph_graphs(),
+///     Some(default.as_slice())
+/// );
 /// # Ok::<_, Box<dyn std::error::Error>>(())
 /// ```
 #[derive(Eq, PartialEq, Debug, Clone, Hash)]
 pub struct Query {
     pub(super) inner: spargebra::Query,
     pub(super) dataset: QueryDataset,
-    pub(super) parsing_duration: Option<Duration>,
+    pub(super) parsing_duration: Option<DayTimeDuration>,
 }
 
 impl Query {
     /// Parses a SPARQL query with an optional base IRI to resolve relative IRIs in the query.
-    pub fn parse(query: &str, base_iri: Option<&str>) -> Result<Self, spargebra::ParseError> {
+    pub fn parse(
+        query: &str,
+        base_iri: Option<&str>,
+    ) -> Result<Self, spargebra::SparqlSyntaxError> {
         let start = Timer::now();
         let query = Self::from(spargebra::Query::parse(query, base_iri)?);
         Ok(Self {
             dataset: query.dataset,
             inner: query.inner,
-            parsing_duration: Some(start.elapsed()),
+            parsing_duration: start.elapsed(),
         })
     }
 
@@ -60,30 +64,30 @@ impl Query {
 
 impl fmt::Display for Query {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        self.inner.fmt(f) //TODO: override
+        self.inner.fmt(f) // TODO: override
     }
 }
 
 impl FromStr for Query {
-    type Err = spargebra::ParseError;
+    type Err = spargebra::SparqlSyntaxError;
 
-    fn from_str(query: &str) -> Result<Self, spargebra::ParseError> {
+    fn from_str(query: &str) -> Result<Self, Self::Err> {
         Self::parse(query, None)
     }
 }
 
-impl<'a> TryFrom<&'a str> for Query {
-    type Error = spargebra::ParseError;
+impl TryFrom<&str> for Query {
+    type Error = spargebra::SparqlSyntaxError;
 
-    fn try_from(query: &str) -> Result<Self, spargebra::ParseError> {
+    fn try_from(query: &str) -> Result<Self, Self::Error> {
         Self::from_str(query)
     }
 }
 
-impl<'a> TryFrom<&'a String> for Query {
-    type Error = spargebra::ParseError;
+impl TryFrom<&String> for Query {
+    type Error = spargebra::SparqlSyntaxError;
 
-    fn try_from(query: &String) -> Result<Self, spargebra::ParseError> {
+    fn try_from(query: &String) -> Result<Self, Self::Error> {
         Self::from_str(query)
     }
 }
@@ -112,7 +116,7 @@ impl From<spargebra::Query> for Query {
 /// let update = Update::parse(update_str, None)?;
 ///
 /// assert_eq!(update.to_string().trim(), update_str);
-/// # Ok::<_, oxigraph::sparql::ParseError>(())
+/// # Ok::<_, oxigraph::sparql::SparqlSyntaxError>(())
 /// ```
 #[derive(Eq, PartialEq, Debug, Clone, Hash)]
 pub struct Update {
@@ -122,22 +126,11 @@ pub struct Update {
 
 impl Update {
     /// Parses a SPARQL update with an optional base IRI to resolve relative IRIs in the query.
-    pub fn parse(update: &str, base_iri: Option<&str>) -> Result<Self, spargebra::ParseError> {
-        let update = spargebra::Update::parse(update, base_iri)?;
-        Ok(Self {
-            using_datasets: update
-                .operations
-                .iter()
-                .map(|operation| {
-                    if let GraphUpdateOperation::DeleteInsert { using, .. } = operation {
-                        Some(QueryDataset::from_algebra(using))
-                    } else {
-                        None
-                    }
-                })
-                .collect(),
-            inner: update,
-        })
+    pub fn parse(
+        update: &str,
+        base_iri: Option<&str>,
+    ) -> Result<Self, spargebra::SparqlSyntaxError> {
+        Ok(spargebra::Update::parse(update, base_iri)?.into())
     }
 
     /// Returns [the query dataset specification](https://www.w3.org/TR/sparql11-query/#specifyingDataset) in [DELETE/INSERT operations](https://www.w3.org/TR/sparql11-update/#deleteInsert).
@@ -158,29 +151,48 @@ impl fmt::Display for Update {
 }
 
 impl FromStr for Update {
-    type Err = spargebra::ParseError;
+    type Err = spargebra::SparqlSyntaxError;
 
-    fn from_str(update: &str) -> Result<Self, spargebra::ParseError> {
+    fn from_str(update: &str) -> Result<Self, Self::Err> {
         Self::parse(update, None)
     }
 }
 
-impl<'a> TryFrom<&'a str> for Update {
-    type Error = spargebra::ParseError;
+impl TryFrom<&str> for Update {
+    type Error = spargebra::SparqlSyntaxError;
 
-    fn try_from(update: &str) -> Result<Self, spargebra::ParseError> {
+    fn try_from(update: &str) -> Result<Self, Self::Error> {
         Self::from_str(update)
     }
 }
 
-impl<'a> TryFrom<&'a String> for Update {
-    type Error = spargebra::ParseError;
+impl TryFrom<&String> for Update {
+    type Error = spargebra::SparqlSyntaxError;
 
-    fn try_from(update: &String) -> Result<Self, spargebra::ParseError> {
+    fn try_from(update: &String) -> Result<Self, Self::Error> {
         Self::from_str(update)
     }
 }
 
+impl From<spargebra::Update> for Update {
+    fn from(update: spargebra::Update) -> Self {
+        Self {
+            using_datasets: update
+                .operations
+                .iter()
+                .map(|operation| {
+                    if let GraphUpdateOperation::DeleteInsert { using, .. } = operation {
+                        Some(QueryDataset::from_algebra(using))
+                    } else {
+                        None
+                    }
+                })
+                .collect(),
+            inner: update,
+        }
+    }
+}
+
 /// A SPARQL query [dataset specification](https://www.w3.org/TR/sparql11-query/#specifyingDataset)
 #[derive(Eq, PartialEq, Debug, Clone, Hash)]
 pub struct QueryDataset {
@@ -219,8 +231,15 @@ impl QueryDataset {
     /// ```
     /// use oxigraph::sparql::Query;
     ///
-    /// assert!(Query::parse("SELECT ?s ?p ?o WHERE { ?s ?p ?o . }", None)?.dataset().is_default_dataset());
-    /// assert!(!Query::parse("SELECT ?s ?p ?o FROM <http://example.com> WHERE { ?s ?p ?o . }", None)?.dataset().is_default_dataset());
+    /// assert!(Query::parse("SELECT ?s ?p ?o WHERE { ?s ?p ?o . }", None)?
+    ///     .dataset()
+    ///     .is_default_dataset());
+    /// assert!(!Query::parse(
+    ///     "SELECT ?s ?p ?o FROM <http://example.com> WHERE { ?s ?p ?o . }",
+    ///     None
+    /// )?
+    /// .dataset()
+    /// .is_default_dataset());
     ///
     /// # Ok::<_, Box<dyn std::error::Error>>(())
     /// ```
@@ -252,7 +271,10 @@ impl QueryDataset {
     /// let mut query = Query::parse("SELECT ?s ?p ?o WHERE { ?s ?p ?o . }", None)?;
     /// let default = vec![NamedNode::new("http://example.com")?.into()];
     /// query.dataset_mut().set_default_graph(default.clone());
-    /// assert_eq!(query.dataset().default_graph_graphs(), Some(default.as_slice()));    
+    /// assert_eq!(
+    ///     query.dataset().default_graph_graphs(),
+    ///     Some(default.as_slice())
+    /// );
     ///
     /// # Ok::<_, Box<dyn std::error::Error>>(())
     /// ```
@@ -273,8 +295,13 @@ impl QueryDataset {
     ///
     /// let mut query = Query::parse("SELECT ?s ?p ?o WHERE { ?s ?p ?o . }", None)?;
     /// let named = vec![NamedNode::new("http://example.com")?.into()];
-    /// query.dataset_mut().set_available_named_graphs(named.clone());
-    /// assert_eq!(query.dataset().available_named_graphs(), Some(named.as_slice()));
+    /// query
+    ///     .dataset_mut()
+    ///     .set_available_named_graphs(named.clone());
+    /// assert_eq!(
+    ///     query.dataset().available_named_graphs(),
+    ///     Some(named.as_slice())
+    /// );
     ///
     /// # Ok::<_, Box<dyn std::error::Error>>(())
     /// ```

lib/oxigraph/src/sparql/dataset.rs

@@ -176,10 +176,6 @@ impl StrLookup for DatasetView {
             self.reader.get_str(key)?
         })
     }
-
-    fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError> {
-        Ok(self.extra.borrow().contains_key(key) || self.reader.contains_str(key)?)
-    }
 }
 
 struct EncodedDatasetSpec {

lib/oxigraph/src/sparql/error.rs

@@ -0,0 +1,84 @@
+use crate::io::RdfParseError;
+use crate::model::NamedNode;
+use crate::sparql::results::QueryResultsParseError as ResultsParseError;
+use crate::sparql::SparqlSyntaxError;
+use crate::storage::StorageError;
+use std::convert::Infallible;
+use std::error::Error;
+use std::io;
+
+/// A SPARQL evaluation error.
+#[derive(Debug, thiserror::Error)]
+#[non_exhaustive]
+pub enum EvaluationError {
+    /// An error in SPARQL parsing.
+    #[error(transparent)]
+    Parsing(#[from] SparqlSyntaxError),
+    /// An error from the storage.
+    #[error(transparent)]
+    Storage(#[from] StorageError),
+    /// An error while parsing an external RDF file.
+    #[error(transparent)]
+    GraphParsing(#[from] RdfParseError),
+    /// An error while parsing an external result file (likely from a federated query).
+    #[error(transparent)]
+    ResultsParsing(#[from] ResultsParseError),
+    /// An error returned during results serialization.
+    #[error(transparent)]
+    ResultsSerialization(#[from] io::Error),
+    /// Error during `SERVICE` evaluation
+    #[error("{0}")]
+    Service(#[source] Box<dyn Error + Send + Sync + 'static>),
+    /// Error when `CREATE` tries to create an already existing graph
+    #[error("The graph {0} already exists")]
+    GraphAlreadyExists(NamedNode),
+    /// Error when `DROP` or `CLEAR` tries to remove a not existing graph
+    #[error("The graph {0} does not exist")]
+    GraphDoesNotExist(NamedNode),
+    /// The variable storing the `SERVICE` name is unbound
+    #[error("The variable encoding the service name is unbound")]
+    UnboundService,
+    /// The given `SERVICE` is not supported
+    #[error("The service {0} is not supported")]
+    UnsupportedService(NamedNode),
+    /// The given content media type returned from an HTTP response is not supported (`SERVICE` and `LOAD`)
+    #[error("The content media type {0} is not supported")]
+    UnsupportedContentType(String),
+    /// The `SERVICE` call has not returns solutions
+    #[error("The service is not returning solutions but a boolean or a graph")]
+    ServiceDoesNotReturnSolutions,
+    /// The results are not a RDF graph
+    #[error("The query results are not a RDF graph")]
+    NotAGraph,
+}
+
+impl From<Infallible> for EvaluationError {
+    #[inline]
+    fn from(error: Infallible) -> Self {
+        match error {}
+    }
+}
+
+impl From<EvaluationError> for io::Error {
+    #[inline]
+    fn from(error: EvaluationError) -> Self {
+        match error {
+            EvaluationError::Parsing(error) => Self::new(io::ErrorKind::InvalidData, error),
+            EvaluationError::GraphParsing(error) => error.into(),
+            EvaluationError::ResultsParsing(error) => error.into(),
+            EvaluationError::ResultsSerialization(error) => error,
+            EvaluationError::Storage(error) => error.into(),
+            EvaluationError::Service(error) => match error.downcast() {
+                Ok(error) => *error,
+                Err(error) => Self::new(io::ErrorKind::Other, error),
+            },
+            EvaluationError::GraphAlreadyExists(_)
+            | EvaluationError::GraphDoesNotExist(_)
+            | EvaluationError::UnboundService
+            | EvaluationError::UnsupportedService(_)
+            | EvaluationError::UnsupportedContentType(_)
+            | EvaluationError::ServiceDoesNotReturnSolutions
+            | EvaluationError::NotAGraph => Self::new(io::ErrorKind::InvalidInput, error),
+        }
+    }
+}

lib/oxigraph/src/sparql/http/dummy.rs

@@ -3,18 +3,18 @@
 use std::io::{Empty, Error, ErrorKind, Result};
 use std::time::Duration;
 
-pub struct Client {}
+pub struct Client;
 
 impl Client {
     pub fn new(_timeout: Option<Duration>, _redirection_limit: usize) -> Self {
-        Self {}
+        Self
     }
 
     #[allow(clippy::unused_self)]
-    pub fn get(&self, _url: &str, _accept: &str) -> Result<(String, Empty)> {
+    pub fn get(&self, _url: &str, _accept: &'static str) -> Result<(String, Empty)> {
         Err(Error::new(
             ErrorKind::Unsupported,
-            "HTTP client is not available. Enable the feature 'http_client'",
+            "HTTP client is not available. Enable the feature 'http-client'",
         ))
     }
 
@@ -23,12 +23,12 @@ impl Client {
         &self,
         _url: &str,
         _payload: Vec<u8>,
-        _content_type: &str,
-        _accept: &str,
+        _content_type: &'static str,
+        _accept: &'static str,
     ) -> Result<(String, Empty)> {
         Err(Error::new(
             ErrorKind::Unsupported,
-            "HTTP client is not available. Enable the feature 'http_client'",
+            "HTTP client is not available. Enable the feature 'http-client'",
         ))
     }
 }

lib/oxigraph/src/sparql/http/mod.rs

@@ -0,0 +1,9 @@
+#[cfg(not(feature = "http-client"))]
+mod dummy;
+#[cfg(feature = "http-client")]
+mod simple;
+
+#[cfg(not(feature = "http-client"))]
+pub use dummy::Client;
+#[cfg(feature = "http-client")]
+pub use simple::Client;

lib/oxigraph/src/sparql/http/simple.rs

@@ -8,18 +8,17 @@ pub struct Client {
 
 impl Client {
     pub fn new(timeout: Option<Duration>, redirection_limit: usize) -> Self {
-        let mut client = oxhttp::Client::new();
+        let mut client = oxhttp::Client::new()
+            .with_redirection_limit(redirection_limit)
+            .with_user_agent(concat!("Oxigraph/", env!("CARGO_PKG_VERSION")))
+            .unwrap();
         if let Some(timeout) = timeout {
-            client.set_global_timeout(timeout);
+            client = client.with_global_timeout(timeout);
         }
-        client.set_redirection_limit(redirection_limit);
-        client
-            .set_user_agent(concat!("Oxigraph/", env!("CARGO_PKG_VERSION")))
-            .unwrap();
         Self { client }
     }
 
-    pub fn get(&self, url: &str, accept: &str) -> Result<(String, Body)> {
+    pub fn get(&self, url: &str, accept: &'static str) -> Result<(String, Body)> {
         let request = Request::builder(Method::GET, url.parse().map_err(invalid_input_error)?)
             .with_header(HeaderName::ACCEPT, accept)
             .map_err(invalid_input_error)?
@@ -50,8 +49,8 @@ impl Client {
         &self,
         url: &str,
         payload: Vec<u8>,
-        content_type: &str,
-        accept: &str,
+        content_type: &'static str,
+        accept: &'static str,
     ) -> Result<(String, Body)> {
         let request = Request::builder(Method::POST, url.parse().map_err(invalid_input_error)?)
             .with_header(HeaderName::ACCEPT, accept)

lib/oxigraph/src/sparql/mod.rs

@@ -8,29 +8,29 @@ mod error;
 mod eval;
 mod http;
 mod model;
-mod plan;
-mod plan_builder;
+pub mod results;
 mod service;
 mod update;
 
 use crate::model::{NamedNode, Term};
 pub use crate::sparql::algebra::{Query, QueryDataset, Update};
 use crate::sparql::dataset::DatasetView;
-pub use crate::sparql::error::{EvaluationError, QueryError};
-use crate::sparql::eval::{SimpleEvaluator, Timer};
+pub use crate::sparql::error::EvaluationError;
+use crate::sparql::eval::{EvalNodeWithStats, SimpleEvaluator, Timer};
 pub use crate::sparql::model::{QueryResults, QuerySolution, QuerySolutionIter, QueryTripleIter};
-use crate::sparql::plan::PlanNodeWithStats;
-use crate::sparql::plan_builder::PlanBuilder;
 pub use crate::sparql::service::ServiceHandler;
 use crate::sparql::service::{EmptyServiceHandler, ErrorConversionServiceHandler};
 pub(crate) use crate::sparql::update::evaluate_update;
 use crate::storage::StorageReader;
-use json_event_parser::{JsonEvent, JsonWriter};
+use json_event_parser::{JsonEvent, ToWriteJsonWriter};
 pub use oxrdf::{Variable, VariableNameParseError};
-pub use sparesults::QueryResultsFormat;
-pub use spargebra::ParseError;
+use oxsdatatypes::{DayTimeDuration, Float};
+pub use spargebra::SparqlSyntaxError;
+use sparopt::algebra::GraphPattern;
+use sparopt::Optimizer;
 use std::collections::HashMap;
 use std::rc::Rc;
+use std::sync::Arc;
 use std::time::Duration;
 use std::{fmt, io};
 
@@ -48,43 +48,39 @@ pub(crate) fn evaluate_query(
         spargebra::Query::Select {
             pattern, base_iri, ..
         } => {
-            let (plan, variables) = PlanBuilder::build(
-                &dataset,
-                &pattern,
-                true,
-                &options.custom_functions,
-                options.without_optimizations,
-            )?;
+            let mut pattern = GraphPattern::from(&pattern);
+            if !options.without_optimizations {
+                pattern = Optimizer::optimize_graph_pattern(pattern);
+            }
             let planning_duration = start_planning.elapsed();
             let (results, explanation) = SimpleEvaluator::new(
                 Rc::new(dataset),
                 base_iri.map(Rc::new),
                 options.service_handler(),
-                Rc::new(options.custom_functions),
+                Arc::new(options.custom_functions),
                 run_stats,
             )
-            .evaluate_select_plan(Rc::new(plan), Rc::new(variables));
+            .evaluate_select(&pattern);
             (Ok(results), explanation, planning_duration)
         }
         spargebra::Query::Ask {
             pattern, base_iri, ..
         } => {
-            let (plan, _) = PlanBuilder::build(
-                &dataset,
-                &pattern,
-                false,
-                &options.custom_functions,
-                options.without_optimizations,
-            )?;
+            let mut pattern = GraphPattern::from(&pattern);
+            if !options.without_optimizations {
+                pattern = Optimizer::optimize_graph_pattern(GraphPattern::Reduced {
+                    inner: Box::new(pattern),
+                });
+            }
             let planning_duration = start_planning.elapsed();
             let (results, explanation) = SimpleEvaluator::new(
                 Rc::new(dataset),
                 base_iri.map(Rc::new),
                 options.service_handler(),
-                Rc::new(options.custom_functions),
+                Arc::new(options.custom_functions),
                 run_stats,
             )
-            .evaluate_ask_plan(Rc::new(plan));
+            .evaluate_ask(&pattern);
             (results, explanation, planning_duration)
         }
         spargebra::Query::Construct {
@@ -93,50 +89,41 @@ pub(crate) fn evaluate_query(
             base_iri,
             ..
         } => {
-            let (plan, variables) = PlanBuilder::build(
-                &dataset,
-                &pattern,
-                false,
-                &options.custom_functions,
-                options.without_optimizations,
-            )?;
-            let construct = PlanBuilder::build_graph_template(
-                &dataset,
-                &template,
-                variables,
-                &options.custom_functions,
-                options.without_optimizations,
-            );
+            let mut pattern = GraphPattern::from(&pattern);
+            if !options.without_optimizations {
+                pattern = Optimizer::optimize_graph_pattern(GraphPattern::Reduced {
+                    inner: Box::new(pattern),
+                });
+            }
             let planning_duration = start_planning.elapsed();
             let (results, explanation) = SimpleEvaluator::new(
                 Rc::new(dataset),
                 base_iri.map(Rc::new),
                 options.service_handler(),
-                Rc::new(options.custom_functions),
+                Arc::new(options.custom_functions),
                 run_stats,
             )
-            .evaluate_construct_plan(Rc::new(plan), construct);
+            .evaluate_construct(&pattern, &template);
             (Ok(results), explanation, planning_duration)
         }
         spargebra::Query::Describe {
             pattern, base_iri, ..
         } => {
-            let (plan, _) = PlanBuilder::build(
-                &dataset,
-                &pattern,
-                false,
-                &options.custom_functions,
-                options.without_optimizations,
-            )?;
+            let mut pattern = GraphPattern::from(&pattern);
+            if !options.without_optimizations {
+                pattern = Optimizer::optimize_graph_pattern(GraphPattern::Reduced {
+                    inner: Box::new(pattern),
+                });
+            }
             let planning_duration = start_planning.elapsed();
             let (results, explanation) = SimpleEvaluator::new(
                 Rc::new(dataset),
                 base_iri.map(Rc::new),
                 options.service_handler(),
-                Rc::new(options.custom_functions),
+                Arc::new(options.custom_functions),
                 run_stats,
             )
-            .evaluate_describe_plan(Rc::new(plan));
+            .evaluate_describe(&pattern);
             (Ok(results), explanation, planning_duration)
         }
     };
@@ -152,36 +139,39 @@ pub(crate) fn evaluate_query(
 /// Options for SPARQL query evaluation.
 ///
 ///
-/// If the `"http_client"` optional feature is enabled,
+/// If the `"http-client"` optional feature is enabled,
 /// a simple HTTP 1.1 client is used to execute [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/) SERVICE calls.
 ///
 /// Usage example disabling the federated query support:
 /// ```
-/// use oxigraph::store::Store;
 /// use oxigraph::sparql::QueryOptions;
+/// use oxigraph::store::Store;
 ///
 /// let store = Store::new()?;
 /// store.query_opt(
 ///     "SELECT * WHERE { SERVICE <https://query.wikidata.org/sparql> {} }",
-///     QueryOptions::default().without_service_handler()
+///     QueryOptions::default().without_service_handler(),
 /// )?;
 /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
 /// ```
 #[derive(Clone, Default)]
 pub struct QueryOptions {
-    service_handler: Option<Rc<dyn ServiceHandler<Error = EvaluationError>>>,
-    custom_functions: HashMap<NamedNode, Rc<dyn Fn(&[Term]) -> Option<Term>>>,
+    service_handler: Option<Arc<dyn ServiceHandler<Error = EvaluationError>>>,
+    custom_functions: CustomFunctionRegistry,
     http_timeout: Option<Duration>,
     http_redirection_limit: usize,
     without_optimizations: bool,
 }
 
+pub(crate) type CustomFunctionRegistry =
+    HashMap<NamedNode, Arc<dyn (Fn(&[Term]) -> Option<Term>) + Send + Sync>>;
+
 impl QueryOptions {
     /// Use a given [`ServiceHandler`] to execute [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/) SERVICE calls.
     #[inline]
     #[must_use]
     pub fn with_service_handler(mut self, service_handler: impl ServiceHandler + 'static) -> Self {
-        self.service_handler = Some(Rc::new(ErrorConversionServiceHandler::wrap(
+        self.service_handler = Some(Arc::new(ErrorConversionServiceHandler::wrap(
             service_handler,
         )));
         self
@@ -191,12 +181,12 @@ impl QueryOptions {
     #[inline]
     #[must_use]
     pub fn without_service_handler(mut self) -> Self {
-        self.service_handler = Some(Rc::new(EmptyServiceHandler));
+        self.service_handler = Some(Arc::new(EmptyServiceHandler));
         self
     }
 
     /// Sets a timeout for HTTP requests done during SPARQL evaluation.
-    #[cfg(feature = "http_client")]
+    #[cfg(feature = "http-client")]
     #[inline]
     #[must_use]
     pub fn with_http_timeout(mut self, timeout: Duration) -> Self {
@@ -207,7 +197,7 @@ impl QueryOptions {
     /// Sets an upper bound of the number of HTTP redirection followed per HTTP request done during SPARQL evaluation.
     ///
     /// By default this value is `0`.
-    #[cfg(feature = "http_client")]
+    #[cfg(feature = "http-client")]
     #[inline]
     #[must_use]
     pub fn with_http_redirection_limit(mut self, redirection_limit: usize) -> Self {
@@ -219,9 +209,9 @@ impl QueryOptions {
     ///
     /// Example with a function serializing terms to N-Triples:
     /// ```
-    /// use oxigraph::store::Store;
     /// use oxigraph::model::*;
     /// use oxigraph::sparql::{QueryOptions, QueryResults};
+    /// use oxigraph::store::Store;
     ///
     /// let store = Store::new()?;
     ///
@@ -229,10 +219,13 @@ impl QueryOptions {
     ///     "SELECT (<http://www.w3.org/ns/formats/N-Triples>(1) AS ?nt) WHERE {}",
     ///     QueryOptions::default().with_custom_function(
     ///         NamedNode::new("http://www.w3.org/ns/formats/N-Triples")?,
-    ///         |args| args.get(0).map(|t| Literal::from(t.to_string()).into())
-    ///     )
+    ///         |args| args.get(0).map(|t| Literal::from(t.to_string()).into()),
+    ///     ),
     /// )? {
-    ///     assert_eq!(solutions.next().unwrap()?.get("nt"), Some(&Literal::from("\"1\"^^<http://www.w3.org/2001/XMLSchema#integer>").into()));
+    ///     assert_eq!(
+    ///         solutions.next().unwrap()?.get("nt"),
+    ///         Some(&Literal::from("\"1\"^^<http://www.w3.org/2001/XMLSchema#integer>").into())
+    ///     );
     /// }
     /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
     /// ```
@@ -241,21 +234,21 @@ impl QueryOptions {
     pub fn with_custom_function(
         mut self,
         name: NamedNode,
-        evaluator: impl Fn(&[Term]) -> Option<Term> + 'static,
+        evaluator: impl Fn(&[Term]) -> Option<Term> + Send + Sync + 'static,
     ) -> Self {
-        self.custom_functions.insert(name, Rc::new(evaluator));
+        self.custom_functions.insert(name, Arc::new(evaluator));
         self
     }
 
-    fn service_handler(&self) -> Rc<dyn ServiceHandler<Error = EvaluationError>> {
+    fn service_handler(&self) -> Arc<dyn ServiceHandler<Error = EvaluationError>> {
         self.service_handler.clone().unwrap_or_else(|| {
-            if cfg!(feature = "http_client") {
-                Rc::new(service::SimpleServiceHandler::new(
+            if cfg!(feature = "http-client") {
+                Arc::new(service::SimpleServiceHandler::new(
                     self.http_timeout,
                     self.http_redirection_limit,
                 ))
             } else {
-                Rc::new(EmptyServiceHandler)
+                Arc::new(EmptyServiceHandler)
             }
         })
     }
@@ -285,28 +278,30 @@ impl From<QueryOptions> for UpdateOptions {
 /// The explanation of a query.
 #[derive(Clone)]
 pub struct QueryExplanation {
-    inner: Rc<PlanNodeWithStats>,
+    inner: Rc<EvalNodeWithStats>,
     with_stats: bool,
-    parsing_duration: Option<Duration>,
-    planning_duration: Duration,
+    parsing_duration: Option<DayTimeDuration>,
+    planning_duration: Option<DayTimeDuration>,
 }
 
 impl QueryExplanation {
     /// Writes the explanation as JSON.
-    pub fn write_in_json(&self, output: impl io::Write) -> io::Result<()> {
-        let mut writer = JsonWriter::from_writer(output);
+    pub fn write_in_json(&self, write: impl io::Write) -> io::Result<()> {
+        let mut writer = ToWriteJsonWriter::new(write);
         writer.write_event(JsonEvent::StartObject)?;
         if let Some(parsing_duration) = self.parsing_duration {
-            writer.write_event(JsonEvent::ObjectKey("parsing duration in seconds"))?;
+            writer.write_event(JsonEvent::ObjectKey("parsing duration in seconds".into()))?;
             writer.write_event(JsonEvent::Number(
-                &parsing_duration.as_secs_f32().to_string(),
+                parsing_duration.as_seconds().to_string().into(),
             ))?;
         }
-        writer.write_event(JsonEvent::ObjectKey("planning duration in seconds"))?;
-        writer.write_event(JsonEvent::Number(
-            &self.planning_duration.as_secs_f32().to_string(),
-        ))?;
-        writer.write_event(JsonEvent::ObjectKey("plan"))?;
+        if let Some(planning_duration) = self.planning_duration {
+            writer.write_event(JsonEvent::ObjectKey("planning duration in seconds".into()))?;
+            writer.write_event(JsonEvent::Number(
+                planning_duration.as_seconds().to_string().into(),
+            ))?;
+        }
+        writer.write_event(JsonEvent::ObjectKey("plan".into()))?;
         self.inner.json_node(&mut writer, self.with_stats)?;
         writer.write_event(JsonEvent::EndObject)
     }
@@ -314,6 +309,20 @@ impl QueryExplanation {
 
 impl fmt::Debug for QueryExplanation {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(f, "{:?}", self.inner)
+        let mut obj = f.debug_struct("QueryExplanation");
+        if let Some(parsing_duration) = self.parsing_duration {
+            obj.field(
+                "parsing duration in seconds",
+                &f32::from(Float::from(parsing_duration.as_seconds())),
+            );
+        }
+        if let Some(planning_duration) = self.planning_duration {
+            obj.field(
+                "planning duration in seconds",
+                &f32::from(Float::from(planning_duration.as_seconds())),
+            );
+        }
+        obj.field("tree", &self.inner);
+        obj.finish_non_exhaustive()
     }
 }

lib/oxigraph/src/sparql/model.rs

@@ -0,0 +1,371 @@
+use crate::io::{RdfFormat, RdfSerializer};
+use crate::model::*;
+use crate::sparql::error::EvaluationError;
+use crate::sparql::results::{
+    FromReadQueryResultsReader, FromReadSolutionsReader, QueryResultsFormat,
+    QueryResultsParseError, QueryResultsParser, QueryResultsSerializer,
+};
+pub use sparesults::QuerySolution;
+use std::io::{Read, Write};
+use std::sync::Arc;
+
+/// Results of a [SPARQL query](https://www.w3.org/TR/sparql11-query/).
+pub enum QueryResults {
+    /// Results of a [SELECT](https://www.w3.org/TR/sparql11-query/#select) query.
+    Solutions(QuerySolutionIter),
+    /// Result of a [ASK](https://www.w3.org/TR/sparql11-query/#ask) query.
+    Boolean(bool),
+    /// Results of a [CONSTRUCT](https://www.w3.org/TR/sparql11-query/#construct) or [DESCRIBE](https://www.w3.org/TR/sparql11-query/#describe) query.
+    Graph(QueryTripleIter),
+}
+
+impl QueryResults {
+    /// Reads a SPARQL query results serialization.
+    pub fn read(
+        read: impl Read + 'static,
+        format: QueryResultsFormat,
+    ) -> Result<Self, QueryResultsParseError> {
+        Ok(QueryResultsParser::from_format(format)
+            .parse_read(read)?
+            .into())
+    }
+
+    /// Writes the query results (solutions or boolean).
+    ///
+    /// This method fails if it is called on the `Graph` results.
+    ///
+    /// ```
+    /// use oxigraph::store::Store;
+    /// use oxigraph::model::*;
+    /// use oxigraph::sparql::results::QueryResultsFormat;
+    ///
+    /// let store = Store::new()?;
+    /// let ex = NamedNodeRef::new("http://example.com")?;
+    /// store.insert(QuadRef::new(ex, ex, ex, GraphNameRef::DefaultGraph))?;
+    ///
+    /// let results = store.query("SELECT ?s WHERE { ?s ?p ?o }")?;
+    /// assert_eq!(
+    ///     results.write(Vec::new(), QueryResultsFormat::Json)?,
+    ///     r#"{"head":{"vars":["s"]},"results":{"bindings":[{"s":{"type":"uri","value":"http://example.com"}}]}}"#.as_bytes()
+    /// );
+    /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+    /// ```
+    pub fn write<W: Write>(
+        self,
+        write: W,
+        format: QueryResultsFormat,
+    ) -> Result<W, EvaluationError> {
+        let serializer = QueryResultsSerializer::from_format(format);
+        match self {
+            Self::Boolean(value) => serializer.serialize_boolean_to_write(write, value),
+            Self::Solutions(solutions) => {
+                let mut writer = serializer
+                    .serialize_solutions_to_write(write, solutions.variables().to_vec())
+                    .map_err(EvaluationError::ResultsSerialization)?;
+                for solution in solutions {
+                    writer
+                        .write(&solution?)
+                        .map_err(EvaluationError::ResultsSerialization)?;
+                }
+                writer.finish()
+            }
+            Self::Graph(triples) => {
+                let s = VariableRef::new_unchecked("subject");
+                let p = VariableRef::new_unchecked("predicate");
+                let o = VariableRef::new_unchecked("object");
+                let mut writer = serializer
+                    .serialize_solutions_to_write(
+                        write,
+                        vec![s.into_owned(), p.into_owned(), o.into_owned()],
+                    )
+                    .map_err(EvaluationError::ResultsSerialization)?;
+                for triple in triples {
+                    let triple = triple?;
+                    writer
+                        .write([
+                            (s, &triple.subject.into()),
+                            (p, &triple.predicate.into()),
+                            (o, &triple.object),
+                        ])
+                        .map_err(EvaluationError::ResultsSerialization)?;
+                }
+                writer.finish()
+            }
+        }
+        .map_err(EvaluationError::ResultsSerialization)
+    }
+
+    /// Writes the graph query results.
+    ///
+    /// This method fails if it is called on the `Solution` or `Boolean` results.
+    ///
+    /// ```
+    /// use oxigraph::io::RdfFormat;
+    /// use oxigraph::model::*;
+    /// use oxigraph::store::Store;
+    ///
+    /// let graph = "<http://example.com> <http://example.com> <http://example.com> .\n";
+    ///
+    /// let store = Store::new()?;
+    /// store.load_graph(
+    ///     graph.as_bytes(),
+    ///     RdfFormat::NTriples,
+    ///     GraphName::DefaultGraph,
+    ///     None,
+    /// )?;
+    ///
+    /// let results = store.query("CONSTRUCT WHERE { ?s ?p ?o }")?;
+    /// assert_eq!(
+    ///     results.write_graph(Vec::new(), RdfFormat::NTriples)?,
+    ///     graph.as_bytes()
+    /// );
+    /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+    /// ```
+    pub fn write_graph<W: Write>(
+        self,
+        write: W,
+        format: impl Into<RdfFormat>,
+    ) -> Result<W, EvaluationError> {
+        if let Self::Graph(triples) = self {
+            let mut writer = RdfSerializer::from_format(format.into()).serialize_to_write(write);
+            for triple in triples {
+                writer
+                    .write_triple(&triple?)
+                    .map_err(EvaluationError::ResultsSerialization)?;
+            }
+            writer
+                .finish()
+                .map_err(EvaluationError::ResultsSerialization)
+        } else {
+            Err(EvaluationError::NotAGraph)
+        }
+    }
+}
+
+impl From<QuerySolutionIter> for QueryResults {
+    #[inline]
+    fn from(value: QuerySolutionIter) -> Self {
+        Self::Solutions(value)
+    }
+}
+
+impl<R: Read + 'static> From<FromReadQueryResultsReader<R>> for QueryResults {
+    fn from(reader: FromReadQueryResultsReader<R>) -> Self {
+        match reader {
+            FromReadQueryResultsReader::Solutions(s) => Self::Solutions(s.into()),
+            FromReadQueryResultsReader::Boolean(v) => Self::Boolean(v),
+        }
+    }
+}
+
+/// An iterator over [`QuerySolution`]s.
+///
+/// ```
+/// use oxigraph::sparql::QueryResults;
+/// use oxigraph::store::Store;
+///
+/// let store = Store::new()?;
+/// if let QueryResults::Solutions(solutions) = store.query("SELECT ?s WHERE { ?s ?p ?o }")? {
+///     for solution in solutions {
+///         println!("{:?}", solution?.get("s"));
+///     }
+/// }
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+pub struct QuerySolutionIter {
+    variables: Arc<[Variable]>,
+    iter: Box<dyn Iterator<Item = Result<QuerySolution, EvaluationError>>>,
+}
+
+impl QuerySolutionIter {
+    /// Construct a new iterator of solution from an ordered list of solution variables and an iterator of solution tuples
+    /// (each tuple using the same ordering as the variable list such that tuple element 0 is the value for the variable 0...)
+    pub fn new(
+        variables: Arc<[Variable]>,
+        iter: impl Iterator<Item = Result<Vec<Option<Term>>, EvaluationError>> + 'static,
+    ) -> Self {
+        Self {
+            variables: Arc::clone(&variables),
+            iter: Box::new(
+                iter.map(move |t| t.map(|values| (Arc::clone(&variables), values).into())),
+            ),
+        }
+    }
+
+    /// The variables used in the solutions.
+    ///
+    /// ```
+    /// use oxigraph::sparql::{QueryResults, Variable};
+    /// use oxigraph::store::Store;
+    ///
+    /// let store = Store::new()?;
+    /// if let QueryResults::Solutions(solutions) = store.query("SELECT ?s ?o WHERE { ?s ?p ?o }")? {
+    ///     assert_eq!(
+    ///         solutions.variables(),
+    ///         &[Variable::new("s")?, Variable::new("o")?]
+    ///     );
+    /// }
+    /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+    /// ```
+    #[inline]
+    pub fn variables(&self) -> &[Variable] {
+        &self.variables
+    }
+}
+
+impl<R: Read + 'static> From<FromReadSolutionsReader<R>> for QuerySolutionIter {
+    fn from(reader: FromReadSolutionsReader<R>) -> Self {
+        Self {
+            variables: reader.variables().into(),
+            iter: Box::new(reader.map(|t| t.map_err(EvaluationError::from))),
+        }
+    }
+}
+
+impl Iterator for QuerySolutionIter {
+    type Item = Result<QuerySolution, EvaluationError>;
+
+    #[inline]
+    fn next(&mut self) -> Option<Self::Item> {
+        self.iter.next()
+    }
+
+    #[inline]
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.iter.size_hint()
+    }
+}
+
+/// An iterator over the triples that compose a graph solution.
+///
+/// ```
+/// use oxigraph::sparql::QueryResults;
+/// use oxigraph::store::Store;
+///
+/// let store = Store::new()?;
+/// if let QueryResults::Graph(triples) = store.query("CONSTRUCT WHERE { ?s ?p ?o }")? {
+///     for triple in triples {
+///         println!("{}", triple?);
+///     }
+/// }
+/// # Result::<_,Box<dyn std::error::Error>>::Ok(())
+/// ```
+pub struct QueryTripleIter {
+    pub(crate) iter: Box<dyn Iterator<Item = Result<Triple, EvaluationError>>>,
+}
+
+impl Iterator for QueryTripleIter {
+    type Item = Result<Triple, EvaluationError>;
+
+    #[inline]
+    fn next(&mut self) -> Option<Self::Item> {
+        self.iter.next()
+    }
+
+    #[inline]
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.iter.size_hint()
+    }
+
+    #[inline]
+    fn fold<Acc, G>(self, init: Acc, g: G) -> Acc
+    where
+        G: FnMut(Acc, Self::Item) -> Acc,
+    {
+        self.iter.fold(init, g)
+    }
+}
+
+#[cfg(test)]
+#[allow(clippy::panic_in_result_fn)]
+mod tests {
+    use super::*;
+    use std::io::Cursor;
+
+    #[test]
+    fn test_serialization_roundtrip() -> Result<(), EvaluationError> {
+        use std::str;
+
+        for format in [
+            QueryResultsFormat::Json,
+            QueryResultsFormat::Xml,
+            QueryResultsFormat::Tsv,
+        ] {
+            let results = vec![
+                QueryResults::Boolean(true),
+                QueryResults::Boolean(false),
+                QueryResults::Solutions(QuerySolutionIter::new(
+                    [
+                        Variable::new_unchecked("foo"),
+                        Variable::new_unchecked("bar"),
+                    ]
+                    .as_ref()
+                    .into(),
+                    Box::new(
+                        vec![
+                            Ok(vec![None, None]),
+                            Ok(vec![
+                                Some(NamedNode::new_unchecked("http://example.com").into()),
+                                None,
+                            ]),
+                            Ok(vec![
+                                None,
+                                Some(NamedNode::new_unchecked("http://example.com").into()),
+                            ]),
+                            Ok(vec![
+                                Some(BlankNode::new_unchecked("foo").into()),
+                                Some(BlankNode::new_unchecked("bar").into()),
+                            ]),
+                            Ok(vec![Some(Literal::new_simple_literal("foo").into()), None]),
+                            Ok(vec![
+                                Some(
+                                    Literal::new_language_tagged_literal_unchecked("foo", "fr")
+                                        .into(),
+                                ),
+                                None,
+                            ]),
+                            Ok(vec![
+                                Some(Literal::from(1).into()),
+                                Some(Literal::from(true).into()),
+                            ]),
+                            Ok(vec![
+                                Some(Literal::from(1.33).into()),
+                                Some(Literal::from(false).into()),
+                            ]),
+                            Ok(vec![
+                                Some(
+                                    Triple::new(
+                                        NamedNode::new_unchecked("http://example.com/s"),
+                                        NamedNode::new_unchecked("http://example.com/p"),
+                                        Triple::new(
+                                            NamedNode::new_unchecked("http://example.com/os"),
+                                            NamedNode::new_unchecked("http://example.com/op"),
+                                            NamedNode::new_unchecked("http://example.com/oo"),
+                                        ),
+                                    )
+                                    .into(),
+                                ),
+                                None,
+                            ]),
+                        ]
+                        .into_iter(),
+                    ),
+                )),
+            ];
+
+            for ex in results {
+                let mut buffer = Vec::new();
+                ex.write(&mut buffer, format)?;
+                let ex2 = QueryResults::read(Cursor::new(buffer.clone()), format)?;
+                let mut buffer2 = Vec::new();
+                ex2.write(&mut buffer2, format)?;
+                assert_eq!(
+                    str::from_utf8(&buffer).unwrap(),
+                    str::from_utf8(&buffer2).unwrap()
+                );
+            }
+        }
+
+        Ok(())
+    }
+}

lib/oxigraph/src/sparql/results.rs

@@ -0,0 +1,44 @@
+//! Utilities to read and write RDF results formats  using [sparesults](https://crates.io/crates/sparesults).
+//!
+//! It supports [SPARQL Query Results XML Format (Second Edition)](https://www.w3.org/TR/rdf-sparql-XMLres/), [SPARQL 1.1 Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/) and [SPARQL 1.1 Query Results CSV and TSV Formats](https://www.w3.org/TR/sparql11-results-csv-tsv/).
+//!
+//! Usage example converting a JSON result file into a TSV result file:
+//!
+//! ```
+//! use oxigraph::sparql::results::{QueryResultsFormat, QueryResultsParser, FromReadQueryResultsReader, QueryResultsSerializer};
+//! use std::io::Result;
+//!
+//! fn convert_json_to_tsv(json_file: &[u8]) -> Result<Vec<u8>> {
+//!     let json_parser = QueryResultsParser::from_format(QueryResultsFormat::Json);
+//!     let tsv_serializer = QueryResultsSerializer::from_format(QueryResultsFormat::Tsv);
+//!     // We start to read the JSON file and see which kind of results it is
+//!     match json_parser.parse_read(json_file)? {
+//!         FromReadQueryResultsReader::Boolean(value) => {
+//!             // it's a boolean result, we copy it in TSV to the output buffer
+//!             tsv_serializer.serialize_boolean_to_write(Vec::new(), value)
+//!         }
+//!         FromReadQueryResultsReader::Solutions(solutions_reader) => {
+//!             // it's a set of solutions, we create a writer and we write to it while reading in streaming from the JSON file
+//!             let mut serialize_solutions_to_write = tsv_serializer.serialize_solutions_to_write(Vec::new(), solutions_reader.variables().to_vec())?;
+//!             for solution in solutions_reader {
+//!                 serialize_solutions_to_write.write(&solution?)?;
+//!             }
+//!             serialize_solutions_to_write.finish()
+//!         }
+//!     }
+//! }
+//!
+//! // Let's test with a boolean
+//! assert_eq!(
+//!     convert_json_to_tsv(br#"{"boolean":true}"#.as_slice()).unwrap(),
+//!     b"true"
+//! );
+//!
+//! // And with a set of solutions
+//! assert_eq!(
+//!     convert_json_to_tsv(br#"{"head":{"vars":["foo","bar"]},"results":{"bindings":[{"foo":{"type":"literal","value":"test"}}]}}"#.as_slice()).unwrap(),
+//!     b"?foo\t?bar\n\"test\"\t\n"
+//! );
+//! ```
+
+pub use sparesults::*;

lib/oxigraph/src/sparql/service.rs

@@ -3,9 +3,8 @@ use crate::sparql::algebra::Query;
 use crate::sparql::error::EvaluationError;
 use crate::sparql::http::Client;
 use crate::sparql::model::QueryResults;
-use crate::sparql::QueryResultsFormat;
+use crate::sparql::results::QueryResultsFormat;
 use std::error::Error;
-use std::io::BufReader;
 use std::time::Duration;
 
 /// Handler for [SPARQL 1.1 Federated Query](https://www.w3.org/TR/sparql11-federated-query/) SERVICE.
@@ -14,18 +13,22 @@ use std::time::Duration;
 /// before evaluating a SPARQL query that uses SERVICE calls.
 ///
 /// ```
-/// use oxigraph::store::Store;
 /// use oxigraph::model::*;
-/// use oxigraph::sparql::{QueryOptions, QueryResults, ServiceHandler, Query, EvaluationError};
+/// use oxigraph::sparql::{EvaluationError, Query, QueryOptions, QueryResults, ServiceHandler};
+/// use oxigraph::store::Store;
 ///
 /// struct TestServiceHandler {
-///     store: Store
+///     store: Store,
 /// }
 ///
 /// impl ServiceHandler for TestServiceHandler {
 ///     type Error = EvaluationError;
 ///
-///     fn handle(&self,service_name: NamedNode, query: Query) -> Result<QueryResults,EvaluationError> {
+///     fn handle(
+///         &self,
+///         service_name: NamedNode,
+///         query: Query,
+///     ) -> Result<QueryResults, Self::Error> {
 ///         if service_name == "http://example.com/service" {
 ///             self.store.query(query)
 ///         } else {
@@ -36,20 +39,23 @@ use std::time::Duration;
 ///
 /// let store = Store::new()?;
 /// let service = TestServiceHandler {
-///     store: Store::new()?
+///     store: Store::new()?,
 /// };
 /// let ex = NamedNodeRef::new("http://example.com")?;
-/// service.store.insert(QuadRef::new(ex, ex, ex, GraphNameRef::DefaultGraph))?;
+/// service
+///     .store
+///     .insert(QuadRef::new(ex, ex, ex, GraphNameRef::DefaultGraph))?;
 ///
 /// if let QueryResults::Solutions(mut solutions) = store.query_opt(
 ///     "SELECT ?s WHERE { SERVICE <http://example.com/service> { ?s ?p ?o } }",
-///     QueryOptions::default().with_service_handler(service)
+///     QueryOptions::default().with_service_handler(service),
 /// )? {
 ///     assert_eq!(solutions.next().unwrap()?.get("s"), Some(&ex.into()));
 /// }
 /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
 /// ```
-pub trait ServiceHandler {
+pub trait ServiceHandler: Send + Sync {
+    /// The service evaluation error.
     type Error: Error + Send + Sync + 'static;
 
     /// Evaluates a [`Query`] against a given service identified by a [`NamedNode`].
@@ -61,10 +67,8 @@ pub struct EmptyServiceHandler;
 impl ServiceHandler for EmptyServiceHandler {
     type Error = EvaluationError;
 
-    fn handle(&self, _: NamedNode, _: Query) -> Result<QueryResults, EvaluationError> {
-        Err(EvaluationError::msg(
-            "The SERVICE feature is not implemented",
-        ))
+    fn handle(&self, name: NamedNode, _: Query) -> Result<QueryResults, Self::Error> {
+        Err(EvaluationError::UnsupportedService(name))
     }
 }
 
@@ -81,14 +85,10 @@ impl<S: ServiceHandler> ErrorConversionServiceHandler<S> {
 impl<S: ServiceHandler> ServiceHandler for ErrorConversionServiceHandler<S> {
     type Error = EvaluationError;
 
-    fn handle(
-        &self,
-        service_name: NamedNode,
-        query: Query,
-    ) -> Result<QueryResults, EvaluationError> {
+    fn handle(&self, service_name: NamedNode, query: Query) -> Result<QueryResults, Self::Error> {
         self.handler
             .handle(service_name, query)
-            .map_err(EvaluationError::wrap)
+            .map_err(|e| EvaluationError::Service(Box::new(e)))
     }
 }
 
@@ -107,22 +107,18 @@ impl SimpleServiceHandler {
 impl ServiceHandler for SimpleServiceHandler {
     type Error = EvaluationError;
 
-    fn handle(
-        &self,
-        service_name: NamedNode,
-        query: Query,
-    ) -> Result<QueryResults, EvaluationError> {
-        let (content_type, body) = self.client.post(
-            service_name.as_str(),
-            query.to_string().into_bytes(),
-            "application/sparql-query",
-            "application/sparql-results+json, application/sparql-results+xml",
-        )?;
-        let format = QueryResultsFormat::from_media_type(&content_type).ok_or_else(|| {
-            EvaluationError::msg(format!(
-                "Unsupported Content-Type returned by {service_name}: {content_type}"
-            ))
-        })?;
-        Ok(QueryResults::read(BufReader::new(body), format)?)
+    fn handle(&self, service_name: NamedNode, query: Query) -> Result<QueryResults, Self::Error> {
+        let (content_type, body) = self
+            .client
+            .post(
+                service_name.as_str(),
+                query.to_string().into_bytes(),
+                "application/sparql-query",
+                "application/sparql-results+json, application/sparql-results+xml",
+            )
+            .map_err(|e| EvaluationError::Service(Box::new(e)))?;
+        let format = QueryResultsFormat::from_media_type(&content_type)
+            .ok_or_else(|| EvaluationError::UnsupportedContentType(content_type))?;
+        Ok(QueryResults::read(body, format)?)
     }
 }

lib/oxigraph/src/sparql/update.rs

@@ -1,12 +1,9 @@
-use crate::io::read::ParseError;
-use crate::io::{GraphFormat, GraphParser};
+use crate::io::{RdfFormat, RdfParser};
 use crate::model::{GraphName as OxGraphName, GraphNameRef, Quad as OxQuad};
 use crate::sparql::algebra::QueryDataset;
 use crate::sparql::dataset::DatasetView;
-use crate::sparql::eval::SimpleEvaluator;
+use crate::sparql::eval::{EncodedTuple, SimpleEvaluator};
 use crate::sparql::http::Client;
-use crate::sparql::plan::EncodedTuple;
-use crate::sparql::plan_builder::PlanBuilder;
 use crate::sparql::{EvaluationError, Update, UpdateOptions};
 use crate::storage::numeric_encoder::{Decoder, EncodedTerm};
 use crate::storage::StorageWriter;
@@ -18,9 +15,11 @@ use spargebra::term::{
     Quad, QuadPattern, Subject, Term, TermPattern, Triple, TriplePattern, Variable,
 };
 use spargebra::GraphUpdateOperation;
+use sparopt::Optimizer;
 use std::collections::HashMap;
-use std::io::BufReader;
+use std::io;
 use std::rc::Rc;
+use std::sync::Arc;
 
 pub fn evaluate_update<'a, 'b: 'a>(
     transaction: &'a mut StorageWriter<'b>,
@@ -74,9 +73,7 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
             } => self.eval_delete_insert(
                 delete,
                 insert,
-                using_dataset
-                    .as_ref()
-                    .ok_or_else(|| EvaluationError::msg("No dataset"))?,
+                using_dataset.as_ref().unwrap_or(&QueryDataset::new()),
                 pattern,
             ),
             GraphUpdateOperation::Load {
@@ -125,24 +122,24 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
         algebra: &GraphPattern,
     ) -> Result<(), EvaluationError> {
         let dataset = Rc::new(DatasetView::new(self.transaction.reader(), using));
-        let (plan, variables) = PlanBuilder::build(
-            &dataset,
-            algebra,
-            false,
-            &self.options.query_options.custom_functions,
-            !self.options.query_options.without_optimizations,
-        )?;
+        let mut pattern = sparopt::algebra::GraphPattern::from(algebra);
+        if !self.options.query_options.without_optimizations {
+            pattern = Optimizer::optimize_graph_pattern(sparopt::algebra::GraphPattern::Reduced {
+                inner: Box::new(pattern),
+            });
+        }
         let evaluator = SimpleEvaluator::new(
             Rc::clone(&dataset),
             self.base_iri.clone(),
             self.options.query_options.service_handler(),
-            Rc::new(self.options.query_options.custom_functions.clone()),
+            Arc::new(self.options.query_options.custom_functions.clone()),
             false,
         );
+        let mut variables = Vec::new();
         let mut bnodes = HashMap::new();
-        let (eval, _) = evaluator.plan_evaluator(Rc::new(plan));
+        let (eval, _) = evaluator.graph_pattern_evaluator(&pattern, &mut variables);
         let tuples =
-            eval(EncodedTuple::with_capacity(variables.len())).collect::<Result<Vec<_>, _>>()?; //TODO: would be much better to stream
+            eval(EncodedTuple::with_capacity(variables.len())).collect::<Result<Vec<_>, _>>()?; // TODO: would be much better to stream
         for tuple in tuples {
             for quad in delete {
                 if let Some(quad) =
@@ -164,28 +161,31 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
     }
 
     fn eval_load(&mut self, from: &NamedNode, to: &GraphName) -> Result<(), EvaluationError> {
-        let (content_type, body) = self.client.get(
-            from.as_str(),
-            "application/n-triples, text/turtle, application/rdf+xml",
-        )?;
-        let format = GraphFormat::from_media_type(&content_type).ok_or_else(|| {
-            EvaluationError::msg(format!(
-                "Unsupported Content-Type returned by {from}: {content_type}"
-            ))
-        })?;
+        let (content_type, body) = self
+            .client
+            .get(
+                from.as_str(),
+                "application/n-triples, text/turtle, application/rdf+xml",
+            )
+            .map_err(|e| EvaluationError::Service(Box::new(e)))?;
+        let format = RdfFormat::from_media_type(&content_type)
+            .ok_or_else(|| EvaluationError::UnsupportedContentType(content_type))?;
         let to_graph_name = match to {
             GraphName::NamedNode(graph_name) => graph_name.into(),
             GraphName::DefaultGraph => GraphNameRef::DefaultGraph,
         };
-        let mut parser = GraphParser::from_format(format);
-        if let Some(base_iri) = &self.base_iri {
-            parser = parser
-                .with_base_iri(base_iri.as_str())
-                .map_err(|e| ParseError::invalid_base_iri(base_iri, e))?;
-        }
-        for t in parser.read_triples(BufReader::new(body))? {
-            self.transaction
-                .insert(t?.as_ref().in_graph(to_graph_name))?;
+        let mut parser = RdfParser::from_format(format)
+            .rename_blank_nodes()
+            .without_named_graphs()
+            .with_default_graph(to_graph_name);
+        parser = parser.with_base_iri(from.as_str()).map_err(|e| {
+            EvaluationError::Service(Box::new(io::Error::new(
+                io::ErrorKind::InvalidInput,
+                format!("Invalid URL: {from}: {e}"),
+            )))
+        })?;
+        for q in parser.parse_read(body) {
+            self.transaction.insert(q?.as_ref())?;
         }
         Ok(())
     }
@@ -194,9 +194,7 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
         if self.transaction.insert_named_graph(graph_name.into())? || silent {
             Ok(())
         } else {
-            Err(EvaluationError::msg(format!(
-                "The graph {graph_name} already exists"
-            )))
+            Err(EvaluationError::GraphAlreadyExists(graph_name.clone()))
         }
     }
 
@@ -212,9 +210,7 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
                 } else if silent {
                     Ok(())
                 } else {
-                    Err(EvaluationError::msg(format!(
-                        "The graph {graph} does not exists"
-                    )))
+                    Err(EvaluationError::GraphDoesNotExist(graph_name.clone()))
                 }
             }
             GraphTarget::DefaultGraph => {
@@ -232,9 +228,7 @@ impl<'a, 'b: 'a> SimpleUpdateEvaluator<'a, 'b> {
                 if self.transaction.remove_named_graph(graph_name.into())? || silent {
                     Ok(())
                 } else {
-                    Err(EvaluationError::msg(format!(
-                        "The graph {graph_name} does not exists"
-                    )))
+                    Err(EvaluationError::GraphDoesNotExist(graph_name.clone()))
                 }
             }
             GraphTarget::DefaultGraph => {

lib/oxigraph/src/storage/backend/fallback.rs

@@ -1,6 +1,7 @@
 //! TODO: This storage is dramatically naive.
 
 use crate::storage::StorageError;
+use crate::store::CorruptionError;
 use std::cell::RefCell;
 use std::collections::{BTreeMap, HashMap};
 use std::error::Error;
@@ -30,9 +31,13 @@ impl Db {
     }
 
     #[allow(clippy::unwrap_in_result)]
-    pub fn column_family(&self, name: &'static str) -> Option<ColumnFamily> {
-        let name = ColumnFamily(name);
-        (self.0.read().unwrap().contains_key(&name)).then(|| name)
+    pub fn column_family(&self, name: &'static str) -> Result<ColumnFamily, StorageError> {
+        let column_family = ColumnFamily(name);
+        if self.0.read().unwrap().contains_key(&column_family) {
+            Ok(column_family)
+        } else {
+            Err(CorruptionError::from_missing_column_family_name(name).into())
+        }
     }
 
     #[must_use]
@@ -116,6 +121,7 @@ impl Reader {
         }
     }
 
+    #[allow(clippy::iter_not_returning_iterator)]
     pub fn iter(&self, column_family: &ColumnFamily) -> Result<Iter, StorageError> {
         self.scan_prefix(column_family, &[])
     }
@@ -129,9 +135,7 @@ impl Reader {
         let data: Vec<_> = match &self.0 {
             InnerReader::Simple(reader) => {
                 let trees = reader.read().unwrap();
-                let tree = if let Some(tree) = trees.get(column_family) {
-                    tree
-                } else {
+                let Some(tree) = trees.get(column_family) else {
                     return Ok(Iter {
                         iter: Vec::new().into_iter(),
                         current: None,
@@ -147,28 +151,25 @@ impl Reader {
                 }
             }
             InnerReader::Transaction(reader) => {
-                if let Some(reader) = reader.upgrade() {
-                    let trees = (*reader).borrow();
-                    let tree = if let Some(tree) = trees.get(column_family) {
-                        tree
-                    } else {
-                        return Ok(Iter {
-                            iter: Vec::new().into_iter(),
-                            current: None,
-                        });
-                    };
-                    if prefix.is_empty() {
-                        tree.iter().map(|(k, v)| (k.clone(), v.clone())).collect()
-                    } else {
-                        tree.range(prefix.to_vec()..)
-                            .take_while(|(k, _)| k.starts_with(prefix))
-                            .map(|(k, v)| (k.clone(), v.clone()))
-                            .collect()
-                    }
-                } else {
+                let Some(reader) = reader.upgrade() else {
                     return Err(StorageError::Other(
                         "The transaction is already ended".into(),
                     ));
+                };
+                let trees = (*reader).borrow();
+                let Some(tree) = trees.get(column_family) else {
+                    return Ok(Iter {
+                        iter: Vec::new().into_iter(),
+                        current: None,
+                    });
+                };
+                if prefix.is_empty() {
+                    tree.iter().map(|(k, v)| (k.clone(), v.clone())).collect()
+                } else {
+                    tree.range(prefix.to_vec()..)
+                        .take_while(|(k, _)| k.starts_with(prefix))
+                        .map(|(k, v)| (k.clone(), v.clone()))
+                        .collect()
                 }
             }
         };
@@ -231,7 +232,7 @@ pub struct Transaction<'a>(
 impl Transaction<'_> {
     #[allow(unsafe_code, clippy::useless_transmute)]
     pub fn reader(&self) -> Reader {
-        // This transmute is safe because we take a weak reference and the only Rc reference used is guarded by the lifetime.
+        // SAFETY: This transmute is safe because we take a weak reference and the only Rc reference used is guarded by the lifetime.
         Reader(InnerReader::Transaction(Rc::downgrade(unsafe {
             transmute(&self.0)
         })))

lib/oxigraph/src/storage/backend/mod.rs

@@ -0,0 +1,12 @@
+//! A storage backend
+//! RocksDB is available, if not in memory
+
+#[cfg(any(target_family = "wasm"))]
+pub use fallback::{ColumnFamily, ColumnFamilyDefinition, Db, Iter, Reader, Transaction};
+#[cfg(all(not(target_family = "wasm")))]
+pub use oxi_rocksdb::{ColumnFamily, ColumnFamilyDefinition, Db, Iter, Reader, Transaction};
+
+#[cfg(any(target_family = "wasm"))]
+mod fallback;
+#[cfg(all(not(target_family = "wasm")))]
+mod oxi_rocksdb;

lib/oxigraph/src/storage/backend/oxi_rocksdb.rs

@@ -1,48 +1,49 @@
 //! Code inspired by [Rust RocksDB](https://github.com/rust-rocksdb/rust-rocksdb) under Apache License 2.0.
 
-#![allow(unsafe_code, trivial_casts)]
+#![allow(
+    unsafe_code,
+    trivial_casts,
+    clippy::undocumented_unsafe_blocks,
+    clippy::panic_in_result_fn,
+    clippy::unwrap_in_result
+)]
 
 use crate::storage::error::{CorruptionError, StorageError};
-use lazy_static::lazy_static;
-use libc::{self, c_void, free};
-use oxrocksdb_sys::*;
+use libc::{c_char, c_void};
 use rand::random;
+use rocksdb::ffi::*;
 use std::borrow::Borrow;
+#[cfg(unix)]
 use std::cmp::min;
 use std::collections::HashMap;
 use std::env::temp_dir;
 use std::error::Error;
 use std::ffi::{CStr, CString};
-use std::fmt;
 use std::fs::remove_dir_all;
-use std::io;
 use std::marker::PhantomData;
 use std::ops::Deref;
 use std::path::{Path, PathBuf};
 use std::rc::{Rc, Weak};
-use std::sync::Arc;
+use std::sync::{Arc, OnceLock};
 use std::thread::{available_parallelism, yield_now};
-use std::{ptr, slice};
-
-macro_rules! ffi_result {
-    ( $($function:ident)::*() ) => {
-        ffi_result_impl!($($function)::*())
-    };
+use std::{fmt, io, ptr, slice};
 
-    ( $($function:ident)::*( $arg1:expr $(, $arg:expr)* $(,)? ) ) => {
-        ffi_result_impl!($($function)::*($arg1 $(, $arg)* ,))
-    };
+pub fn opt_bytes_to_ptr<T: AsRef<[u8]>>(opt: Option<T>) -> *const c_char {
+    match opt {
+        Some(v) => v.as_ref().as_ptr() as *const c_char,
+        None => ptr::null(),
+    }
 }
 
-macro_rules! ffi_result_impl {
-    ( $($function:ident)::*( $($arg:expr,)*) ) => {{
+macro_rules! ffi_result {
+    ( $($function:ident)::*( $arg1:expr $(, $arg:expr)* $(,)? ) ) => {{
         let mut status = rocksdb_status_t {
             code: rocksdb_status_code_t_rocksdb_status_code_ok,
             subcode: rocksdb_status_subcode_t_rocksdb_status_subcode_none,
             severity: rocksdb_status_severity_t_rocksdb_status_severity_none,
             string: ptr::null()
         };
-        let result = $($function)::*($($arg,)* &mut status);
+        let result = $($function)::*($arg1 $(, $arg)* , &mut status);
         if status.code == rocksdb_status_code_t_rocksdb_status_code_ok {
             Ok(result)
         } else {
@@ -51,23 +52,6 @@ macro_rules! ffi_result_impl {
     }}
 }
 
-lazy_static! {
-    static ref ROCKSDB_ENV: UnsafeEnv = {
-        unsafe {
-            let env = rocksdb_create_default_env();
-            assert!(!env.is_null(), "rocksdb_create_default_env returned null");
-            UnsafeEnv(env)
-        }
-    };
-    static ref ROCKSDB_MEM_ENV: UnsafeEnv = {
-        unsafe {
-            let env = rocksdb_create_mem_env();
-            assert!(!env.is_null(), "rocksdb_create_mem_env returned null");
-            UnsafeEnv(env)
-        }
-    };
-}
-
 pub struct ColumnFamilyDefinition {
     pub name: &'static str,
     pub use_iter: bool,
@@ -131,8 +115,7 @@ impl Drop for RwDbHandler {
             rocksdb_block_based_options_destroy(self.block_based_table_options);
         }
         if self.in_memory {
-            #[allow(clippy::let_underscore_must_use)]
-            let _ = remove_dir_all(&self.path);
+            drop(remove_dir_all(&self.path));
         }
     }
 }
@@ -166,20 +149,20 @@ impl Drop for RoDbHandler {
             rocksdb_options_destroy(self.options);
         }
         if let Some(path) = &self.path_to_remove {
-            #[allow(clippy::let_underscore_must_use)]
-            let _ = remove_dir_all(path);
+            drop(remove_dir_all(path));
         }
     }
 }
 
 impl Db {
     pub fn new(column_families: Vec<ColumnFamilyDefinition>) -> Result<Self, StorageError> {
-        Self::open_read_write(None, column_families)
+        Self::open_read_write(None, column_families, None)
     }
 
     pub fn open_read_write(
         path: Option<&Path>,
         column_families: Vec<ColumnFamilyDefinition>,
+        key: Option<[u8; 32]>,
     ) -> Result<Self, StorageError> {
         let (path, in_memory) = if let Some(path) = path {
             (path.to_path_buf(), false)
@@ -188,7 +171,7 @@ impl Db {
         };
         let c_path = path_to_cstring(&path)?;
         unsafe {
-            let options = Self::db_options(true, in_memory)?;
+            let options = Self::db_options(true, in_memory, key)?;
             rocksdb_options_set_create_if_missing(options, 1);
             rocksdb_options_set_create_missing_column_families(options, 1);
             rocksdb_options_set_compression(
@@ -212,7 +195,7 @@ impl Db {
                 16,
             );
             rocksdb_options_set_block_based_table_factory(options, block_based_table_options);
-            #[cfg(feature = "rocksdb_debug")]
+            #[cfg(feature = "rocksdb-debug")]
             {
                 rocksdb_options_set_info_log_level(options, 0);
                 rocksdb_options_enable_statistics(options);
@@ -272,6 +255,7 @@ impl Db {
                 !write_options.is_null(),
                 "rocksdb_writeoptions_create returned null"
             );
+            rocksdb_writeoptions_set_sync(write_options, 1);
             if in_memory {
                 rocksdb_writeoptions_disable_WAL(write_options, 1); // No need for WAL
             }
@@ -330,82 +314,83 @@ impl Db {
         }
     }
 
-    pub fn open_secondary(
-        primary_path: &Path,
-        secondary_path: Option<&Path>,
-        column_families: Vec<ColumnFamilyDefinition>,
-    ) -> Result<Self, StorageError> {
-        let c_primary_path = path_to_cstring(primary_path)?;
-        let (secondary_path, in_memory) = if let Some(path) = secondary_path {
-            (path.to_path_buf(), false)
-        } else {
-            (tmp_path(), true)
-        };
-        let c_secondary_path = path_to_cstring(&secondary_path)?;
-        unsafe {
-            let options = Self::db_options(false, false)?;
-            let (column_family_names, c_column_family_names, cf_options) =
-                Self::column_families_names_and_options(column_families, options);
-            let mut cf_handles: Vec<*mut rocksdb_column_family_handle_t> =
-                vec![ptr::null_mut(); column_family_names.len()];
-            let c_num_column_families = c_column_family_names.len().try_into().unwrap();
-            let db = ffi_result!(rocksdb_open_as_secondary_column_families_with_status(
-                options,
-                c_primary_path.as_ptr(),
-                c_secondary_path.as_ptr(),
-                c_num_column_families,
-                c_column_family_names
-                    .iter()
-                    .map(|cf| cf.as_ptr())
-                    .collect::<Vec<_>>()
-                    .as_ptr(),
-                cf_options.as_ptr().cast(),
-                cf_handles.as_mut_ptr(),
-            ))
-            .map_err(|e| {
-                for cf_option in &cf_options {
-                    rocksdb_options_destroy(*cf_option);
-                }
-                rocksdb_options_destroy(options);
-                e
-            })?;
-            assert!(
-                !db.is_null(),
-                "rocksdb_open_for_read_only_column_families_with_status returned null"
-            );
-            for handle in &cf_handles {
-                assert!(
-                    !handle.is_null(),
-                    "rocksdb_open_for_read_only_column_families_with_status returned a null column family"
-                );
-            }
-            let read_options = rocksdb_readoptions_create();
-            assert!(
-                !read_options.is_null(),
-                "rocksdb_readoptions_create returned null"
-            );
-            Ok(Self {
-                inner: DbKind::ReadOnly(Arc::new(RoDbHandler {
-                    db,
-                    options,
-                    read_options,
-                    column_family_names,
-                    cf_handles,
-                    cf_options,
-                    is_secondary: true,
-                    path_to_remove: in_memory.then(|| secondary_path),
-                })),
-            })
-        }
-    }
+    // pub fn open_secondary(
+    //     primary_path: &Path,
+    //     secondary_path: Option<&Path>,
+    //     column_families: Vec<ColumnFamilyDefinition>,
+    // ) -> Result<Self, StorageError> {
+    //     let c_primary_path = path_to_cstring(primary_path)?;
+    //     let (secondary_path, in_memory) = if let Some(path) = secondary_path {
+    //         (path.to_path_buf(), false)
+    //     } else {
+    //         (tmp_path(), true)
+    //     };
+    //     let c_secondary_path = path_to_cstring(&secondary_path)?;
+    //     unsafe {
+    //         let options = Self::db_options(false, false)?;
+    //         let (column_family_names, c_column_family_names, cf_options) =
+    //             Self::column_families_names_and_options(column_families, options);
+    //         let mut cf_handles: Vec<*mut rocksdb_column_family_handle_t> =
+    //             vec![ptr::null_mut(); column_family_names.len()];
+    //         let c_num_column_families = c_column_family_names.len().try_into().unwrap();
+    //         let db = ffi_result!(rocksdb_open_as_secondary_column_families_with_status(
+    //             options,
+    //             c_primary_path.as_ptr(),
+    //             c_secondary_path.as_ptr(),
+    //             c_num_column_families,
+    //             c_column_family_names
+    //                 .iter()
+    //                 .map(|cf| cf.as_ptr())
+    //                 .collect::<Vec<_>>()
+    //                 .as_ptr(),
+    //             cf_options.as_ptr().cast(),
+    //             cf_handles.as_mut_ptr(),
+    //         ))
+    //         .map_err(|e| {
+    //             for cf_option in &cf_options {
+    //                 rocksdb_options_destroy(*cf_option);
+    //             }
+    //             rocksdb_options_destroy(options);
+    //             e
+    //         })?;
+    //         assert!(
+    //             !db.is_null(),
+    //             "rocksdb_open_for_read_only_column_families_with_status returned null"
+    //         );
+    //         for handle in &cf_handles {
+    //             assert!(
+    //                 !handle.is_null(),
+    //                 "rocksdb_open_for_read_only_column_families_with_status returned a null column family"
+    //             );
+    //         }
+    //         let read_options = rocksdb_readoptions_create();
+    //         assert!(
+    //             !read_options.is_null(),
+    //             "rocksdb_readoptions_create returned null"
+    //         );
+    //         Ok(Self {
+    //             inner: DbKind::ReadOnly(Arc::new(RoDbHandler {
+    //                 db,
+    //                 options,
+    //                 read_options,
+    //                 column_family_names,
+    //                 cf_handles,
+    //                 cf_options,
+    //                 is_secondary: true,
+    //                 path_to_remove: in_memory.then_some(secondary_path),
+    //             })),
+    //         })
+    //     }
+    // }
 
     pub fn open_read_only(
         path: &Path,
         column_families: Vec<ColumnFamilyDefinition>,
+        key: Option<[u8; 32]>,
     ) -> Result<Self, StorageError> {
         unsafe {
             let c_path = path_to_cstring(path)?;
-            let options = Self::db_options(true, false)?;
+            let options = Self::db_options(true, false, key)?;
             let (column_family_names, c_column_family_names, cf_options) =
                 Self::column_families_names_and_options(column_families, options);
             let mut cf_handles: Vec<*mut rocksdb_column_family_handle_t> =
@@ -465,7 +450,11 @@ impl Db {
     fn db_options(
         limit_max_open_files: bool,
         in_memory: bool,
+        key: Option<[u8; 32]>,
     ) -> Result<*mut rocksdb_options_t, StorageError> {
+        static ROCKSDB_ENV: OnceLock<UnsafeEnv> = OnceLock::new();
+        static ROCKSDB_MEM_ENV: OnceLock<UnsafeEnv> = OnceLock::new();
+
         unsafe {
             let options = rocksdb_options_create();
             assert!(!options.is_null(), "rocksdb_options_create returned null");
@@ -502,10 +491,22 @@ impl Db {
             rocksdb_options_set_env(
                 options,
                 if in_memory {
-                    ROCKSDB_MEM_ENV.0
+                    ROCKSDB_MEM_ENV.get_or_init(|| {
+                        let env = rocksdb_create_mem_env();
+                        assert!(!env.is_null(), "rocksdb_create_mem_env returned null");
+                        UnsafeEnv(env)
+                    })
                 } else {
-                    ROCKSDB_ENV.0
-                },
+                    ROCKSDB_ENV.get_or_init(|| {
+                        let env = match key {
+                            Some(_) => rocksdb_create_encrypted_env(opt_bytes_to_ptr(key.as_ref())),
+                            None => rocksdb_create_default_env(),
+                        };
+                        assert!(!env.is_null(), "rocksdb_create_encrypted_env returned null");
+                        UnsafeEnv(env)
+                    })
+                }
+                .0,
             );
             Ok(options)
         }
@@ -551,17 +552,17 @@ impl Db {
         (column_family_names, c_column_family_names, cf_options)
     }
 
-    pub fn column_family(&self, name: &'static str) -> Option<ColumnFamily> {
+    pub fn column_family(&self, name: &'static str) -> Result<ColumnFamily, StorageError> {
         let (column_family_names, cf_handles) = match &self.inner {
             DbKind::ReadOnly(db) => (&db.column_family_names, &db.cf_handles),
             DbKind::ReadWrite(db) => (&db.column_family_names, &db.cf_handles),
         };
         for (cf, cf_handle) in column_family_names.iter().zip(cf_handles) {
             if *cf == name {
-                return Some(ColumnFamily(*cf_handle));
+                return Ok(ColumnFamily(*cf_handle));
             }
         }
-        None
+        Err(CorruptionError::from_missing_column_family_name(name).into())
     }
 
     #[must_use]
@@ -571,8 +572,9 @@ impl Db {
                 DbKind::ReadOnly(db) => {
                     if db.is_secondary {
                         // We try to refresh (and ignore the errors)
-                        #[allow(clippy::let_underscore_must_use)]
-                        let _ = ffi_result!(rocksdb_try_catch_up_with_primary_with_status(db.db));
+                        drop(ffi_result!(rocksdb_try_catch_up_with_primary_with_status(
+                            db.db
+                        )));
                     }
                     let options = rocksdb_readoptions_create_copy(db.read_options);
                     Reader {
@@ -626,7 +628,7 @@ impl Db {
                     (options, snapshot)
                 };
                 let result = f(Transaction {
-                    transaction: Rc::new(transaction),
+                    inner: Rc::new(transaction),
                     read_options,
                     _lifetime: PhantomData,
                 });
@@ -637,7 +639,7 @@ impl Db {
                                 ffi_result!(rocksdb_transaction_commit_with_status(transaction));
                             rocksdb_transaction_destroy(transaction);
                             rocksdb_readoptions_destroy(read_options);
-                            free(snapshot as *mut c_void);
+                            rocksdb_free(snapshot as *mut c_void);
                             r.map_err(StorageError::from)?; // We make sure to also run destructors if the commit fails
                         }
                         return Ok(result);
@@ -648,7 +650,7 @@ impl Db {
                                 ffi_result!(rocksdb_transaction_rollback_with_status(transaction));
                             rocksdb_transaction_destroy(transaction);
                             rocksdb_readoptions_destroy(read_options);
-                            free(snapshot as *mut c_void);
+                            rocksdb_free(snapshot as *mut c_void);
                             r.map_err(StorageError::from)?; // We make sure to also run destructors if the commit fails
                         }
                         // We look for the root error
@@ -721,7 +723,7 @@ impl Db {
         column_family: &ColumnFamily,
         key: &[u8],
     ) -> Result<bool, StorageError> {
-        Ok(self.get(column_family, key)?.is_some()) //TODO: optimize
+        Ok(self.get(column_family, key)?.is_some()) // TODO: optimize
     }
 
     pub fn insert(
@@ -750,13 +752,14 @@ impl Db {
         }
     }
 
-    pub fn flush(&self, column_family: &ColumnFamily) -> Result<(), StorageError> {
+    pub fn flush(&self) -> Result<(), StorageError> {
         if let DbKind::ReadWrite(db) = &self.inner {
             unsafe {
-                ffi_result!(rocksdb_transactiondb_flush_cf_with_status(
+                ffi_result!(rocksdb_transactiondb_flush_cfs_with_status(
                     db.db,
                     db.flush_options,
-                    column_family.0,
+                    db.cf_handles.as_ptr().cast_mut(),
+                    db.cf_handles.len().try_into().unwrap()
                 ))
             }?;
             Ok(())
@@ -944,19 +947,18 @@ impl Reader {
                     ))
                 }
                 InnerReader::Transaction(inner) => {
-                    if let Some(inner) = inner.upgrade() {
-                        ffi_result!(rocksdb_transaction_get_pinned_cf_with_status(
-                            *inner,
-                            self.options,
-                            column_family.0,
-                            key.as_ptr().cast(),
-                            key.len()
-                        ))
-                    } else {
+                    let Some(inner) = inner.upgrade() else {
                         return Err(StorageError::Other(
                             "The transaction is already ended".into(),
                         ));
-                    }
+                    };
+                    ffi_result!(rocksdb_transaction_get_pinned_cf_with_status(
+                        *inner,
+                        self.options,
+                        column_family.0,
+                        key.as_ptr().cast(),
+                        key.len()
+                    ))
                 }
                 InnerReader::PlainDb(inner) => {
                     ffi_result!(rocksdb_get_pinned_cf_with_status(
@@ -981,9 +983,10 @@ impl Reader {
         column_family: &ColumnFamily,
         key: &[u8],
     ) -> Result<bool, StorageError> {
-        Ok(self.get(column_family, key)?.is_some()) //TODO: optimize
+        Ok(self.get(column_family, key)?.is_some()) // TODO: optimize
     }
 
+    #[allow(clippy::iter_not_returning_iterator)]
     pub fn iter(&self, column_family: &ColumnFamily) -> Result<Iter, StorageError> {
         self.scan_prefix(column_family, &[])
     }
@@ -993,7 +996,7 @@ impl Reader {
         column_family: &ColumnFamily,
         prefix: &[u8],
     ) -> Result<Iter, StorageError> {
-        //We generate the upper bound
+        // We generate the upper bound
         let upper_bound = {
             let mut bound = prefix.to_vec();
             let mut found = false;
@@ -1004,7 +1007,7 @@ impl Reader {
                     break;
                 }
             }
-            found.then(|| bound)
+            found.then_some(bound)
         };
 
         unsafe {
@@ -1025,13 +1028,12 @@ impl Reader {
                     rocksdb_transactiondb_create_iterator_cf(inner.db.db, options, column_family.0)
                 }
                 InnerReader::Transaction(inner) => {
-                    if let Some(inner) = inner.upgrade() {
-                        rocksdb_transaction_create_iterator_cf(*inner, options, column_family.0)
-                    } else {
+                    let Some(inner) = inner.upgrade() else {
                         return Err(StorageError::Other(
                             "The transaction is already ended".into(),
                         ));
-                    }
+                    };
+                    rocksdb_transaction_create_iterator_cf(*inner, options, column_family.0)
                 }
                 InnerReader::PlainDb(inner) => {
                     rocksdb_create_iterator_cf(inner.db, options, column_family.0)
@@ -1045,7 +1047,7 @@ impl Reader {
             }
             let is_currently_valid = rocksdb_iter_valid(iter) != 0;
             Ok(Iter {
-                iter,
+                inner: iter,
                 options,
                 _upper_bound: upper_bound,
                 _reader: self.clone(),
@@ -1073,7 +1075,7 @@ impl Reader {
 }
 
 pub struct Transaction<'a> {
-    transaction: Rc<*mut rocksdb_transaction_t>,
+    inner: Rc<*mut rocksdb_transaction_t>,
     read_options: *mut rocksdb_readoptions_t,
     _lifetime: PhantomData<&'a ()>,
 }
@@ -1081,7 +1083,7 @@ pub struct Transaction<'a> {
 impl Transaction<'_> {
     pub fn reader(&self) -> Reader {
         Reader {
-            inner: InnerReader::Transaction(Rc::downgrade(&self.transaction)),
+            inner: InnerReader::Transaction(Rc::downgrade(&self.inner)),
             options: unsafe { rocksdb_readoptions_create_copy(self.read_options) },
         }
     }
@@ -1093,7 +1095,7 @@ impl Transaction<'_> {
     ) -> Result<Option<PinnableSlice>, StorageError> {
         unsafe {
             let slice = ffi_result!(rocksdb_transaction_get_for_update_pinned_cf_with_status(
-                *self.transaction,
+                *self.inner,
                 self.read_options,
                 column_family.0,
                 key.as_ptr().cast(),
@@ -1112,7 +1114,7 @@ impl Transaction<'_> {
         column_family: &ColumnFamily,
         key: &[u8],
     ) -> Result<bool, StorageError> {
-        Ok(self.get_for_update(column_family, key)?.is_some()) //TODO: optimize
+        Ok(self.get_for_update(column_family, key)?.is_some()) // TODO: optimize
     }
 
     pub fn insert(
@@ -1123,7 +1125,7 @@ impl Transaction<'_> {
     ) -> Result<(), StorageError> {
         unsafe {
             ffi_result!(rocksdb_transaction_put_cf_with_status(
-                *self.transaction,
+                *self.inner,
                 column_family.0,
                 key.as_ptr().cast(),
                 key.len(),
@@ -1145,7 +1147,7 @@ impl Transaction<'_> {
     pub fn remove(&mut self, column_family: &ColumnFamily, key: &[u8]) -> Result<(), StorageError> {
         unsafe {
             ffi_result!(rocksdb_transaction_delete_cf_with_status(
-                *self.transaction,
+                *self.inner,
                 column_family.0,
                 key.as_ptr().cast(),
                 key.len(),
@@ -1168,7 +1170,7 @@ impl Drop for PinnableSlice {
 impl Deref for PinnableSlice {
     type Target = [u8];
 
-    fn deref(&self) -> &[u8] {
+    fn deref(&self) -> &Self::Target {
         unsafe {
             let mut len = 0;
             let val = rocksdb_pinnableslice_value(self.0, &mut len);
@@ -1203,7 +1205,7 @@ pub struct Buffer {
 impl Drop for Buffer {
     fn drop(&mut self) {
         unsafe {
-            free(self.base.cast());
+            rocksdb_free(self.base.cast());
         }
     }
 }
@@ -1211,7 +1213,7 @@ impl Drop for Buffer {
 impl Deref for Buffer {
     type Target = [u8];
 
-    fn deref(&self) -> &[u8] {
+    fn deref(&self) -> &Self::Target {
         unsafe { slice::from_raw_parts(self.base, self.len) }
     }
 }
@@ -1235,17 +1237,17 @@ impl From<Buffer> for Vec<u8> {
 }
 
 pub struct Iter {
-    iter: *mut rocksdb_iterator_t,
+    inner: *mut rocksdb_iterator_t,
     is_currently_valid: bool,
     _upper_bound: Option<Vec<u8>>,
     _reader: Reader, // needed to ensure that DB still lives while iter is used
-    options: *mut rocksdb_readoptions_t, // needed to ensure that options still lives while iter is used
+    options: *mut rocksdb_readoptions_t, /* needed to ensure that options still lives while iter is used */
 }
 
 impl Drop for Iter {
     fn drop(&mut self) {
         unsafe {
-            rocksdb_iter_destroy(self.iter);
+            rocksdb_iter_destroy(self.inner);
             rocksdb_readoptions_destroy(self.options);
         }
     }
@@ -1263,15 +1265,15 @@ impl Iter {
 
     pub fn status(&self) -> Result<(), StorageError> {
         unsafe {
-            ffi_result!(rocksdb_iter_get_status(self.iter))?;
+            ffi_result!(rocksdb_iter_get_status(self.inner))?;
         }
         Ok(())
     }
 
     pub fn next(&mut self) {
         unsafe {
-            rocksdb_iter_next(self.iter);
-            self.is_currently_valid = rocksdb_iter_valid(self.iter) != 0;
+            rocksdb_iter_next(self.inner);
+            self.is_currently_valid = rocksdb_iter_valid(self.inner) != 0;
         }
     }
 
@@ -1279,7 +1281,7 @@ impl Iter {
         if self.is_valid() {
             unsafe {
                 let mut len = 0;
-                let val = rocksdb_iter_key(self.iter, &mut len);
+                let val = rocksdb_iter_key(self.inner, &mut len);
                 Some(slice::from_raw_parts(val.cast(), len))
             }
         } else {
@@ -1327,6 +1329,8 @@ impl SstFileWriter {
     }
 }
 
+#[derive(thiserror::Error)]
+#[error("{}", self.message())]
 struct ErrorStatus(rocksdb_status_t);
 
 unsafe impl Send for ErrorStatus {}
@@ -1336,7 +1340,7 @@ impl Drop for ErrorStatus {
     fn drop(&mut self) {
         if !self.0.string.is_null() {
             unsafe {
-                free(self.0.string as *mut c_void);
+                rocksdb_free(self.0.string as *mut c_void);
             }
         }
     }
@@ -1365,14 +1369,6 @@ impl fmt::Debug for ErrorStatus {
     }
 }
 
-impl fmt::Display for ErrorStatus {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        f.write_str(self.message())
-    }
-}
-
-impl Error for ErrorStatus {}
-
 impl From<ErrorStatus> for StorageError {
     fn from(status: ErrorStatus) -> Self {
         if status.0.code == rocksdb_status_code_t_rocksdb_status_code_io_error {
@@ -1397,7 +1393,7 @@ impl From<ErrorStatus> for StorageError {
 
 struct UnsafeEnv(*mut rocksdb_env_t);
 
-// Hack for lazy_static. OK because only written in lazy static and used in a thread-safe way by RocksDB
+// Hack for OnceCell. OK because only written in OnceCell and used in a thread-safe way by RocksDB
 unsafe impl Send for UnsafeEnv {}
 unsafe impl Sync for UnsafeEnv {}
 
@@ -1417,7 +1413,7 @@ fn path_to_cstring(path: &Path) -> Result<CString, StorageError> {
 }
 
 #[cfg(unix)]
-fn available_file_descriptors() -> io::Result<Option<u64>> {
+fn available_file_descriptors() -> io::Result<Option<libc::rlim_t>> {
     let mut rlimit = libc::rlimit {
         rlim_cur: 0,
         rlim_max: 0,
@@ -1430,12 +1426,12 @@ fn available_file_descriptors() -> io::Result<Option<u64>> {
 }
 
 #[cfg(windows)]
-fn available_file_descriptors() -> io::Result<Option<u64>> {
+fn available_file_descriptors() -> io::Result<Option<libc::c_int>> {
     Ok(Some(512)) // https://docs.microsoft.com/en-us/cpp/c-runtime-library/file-handling
 }
 
 #[cfg(not(any(unix, windows)))]
-fn available_file_descriptors() -> io::Result<Option<u64>> {
+fn available_file_descriptors() -> io::Result<Option<libc::c_int>> {
     Ok(None)
 }
 

lib/oxigraph/src/storage/binary_encoder.rs

@@ -2,10 +2,10 @@ use crate::storage::error::{CorruptionError, StorageError};
 use crate::storage::numeric_encoder::{EncodedQuad, EncodedTerm, EncodedTriple, StrHash};
 use crate::storage::small_string::SmallString;
 use oxsdatatypes::*;
-use std::io::{Cursor, Read};
+use std::io::Read;
 use std::mem::size_of;
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 pub const LATEST_STORAGE_VERSION: u64 = 1;
 pub const WRITTEN_TERM_MAX_SIZE: usize = size_of::<u8>() + 2 * size_of::<StrHash>();
 
@@ -62,24 +62,23 @@ pub enum QuadEncoding {
 }
 
 impl QuadEncoding {
-    pub fn decode(self, buffer: &[u8]) -> Result<EncodedQuad, StorageError> {
-        let mut cursor = Cursor::new(&buffer);
+    pub fn decode(self, mut buffer: &[u8]) -> Result<EncodedQuad, StorageError> {
         match self {
-            Self::Spog => cursor.read_spog_quad(),
-            Self::Posg => cursor.read_posg_quad(),
-            Self::Ospg => cursor.read_ospg_quad(),
-            Self::Gspo => cursor.read_gspo_quad(),
-            Self::Gpos => cursor.read_gpos_quad(),
-            Self::Gosp => cursor.read_gosp_quad(),
-            Self::Dspo => cursor.read_dspo_quad(),
-            Self::Dpos => cursor.read_dpos_quad(),
-            Self::Dosp => cursor.read_dosp_quad(),
+            Self::Spog => buffer.read_spog_quad(),
+            Self::Posg => buffer.read_posg_quad(),
+            Self::Ospg => buffer.read_ospg_quad(),
+            Self::Gspo => buffer.read_gspo_quad(),
+            Self::Gpos => buffer.read_gpos_quad(),
+            Self::Gosp => buffer.read_gosp_quad(),
+            Self::Dspo => buffer.read_dspo_quad(),
+            Self::Dpos => buffer.read_dpos_quad(),
+            Self::Dosp => buffer.read_dosp_quad(),
         }
     }
 }
 
-pub fn decode_term(buffer: &[u8]) -> Result<EncodedTerm, StorageError> {
-    Cursor::new(&buffer).read_term()
+pub fn decode_term(mut buffer: &[u8]) -> Result<EncodedTerm, StorageError> {
+    buffer.read_term()
 }
 
 pub trait TermReader {
@@ -633,6 +632,7 @@ pub fn write_term(sink: &mut Vec<u8>, term: &EncodedTerm) {
 }
 
 #[cfg(test)]
+#[allow(clippy::panic_in_result_fn)]
 mod tests {
     use super::*;
     use crate::model::TermRef;
@@ -649,10 +649,6 @@ mod tests {
         fn get_str(&self, key: &StrHash) -> Result<Option<String>, StorageError> {
             Ok(self.id2str.borrow().get(key).cloned())
         }
-
-        fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError> {
-            Ok(self.id2str.borrow().contains_key(key))
-        }
     }
 
     impl MemoryStrStore {
@@ -740,7 +736,7 @@ mod tests {
 
             let mut buffer = Vec::new();
             write_term(&mut buffer, &encoded);
-            assert_eq!(encoded, Cursor::new(&buffer).read_term().unwrap());
+            assert_eq!(encoded, buffer.as_slice().read_term().unwrap());
         }
     }
 }

lib/oxigraph/src/storage/error.rs

@@ -0,0 +1,139 @@
+use crate::io::{RdfFormat, RdfParseError};
+use crate::storage::numeric_encoder::EncodedTerm;
+use oxiri::IriParseError;
+use oxrdf::TermRef;
+use std::error::Error;
+use std::io;
+
+/// An error related to storage operations (reads, writes...).
+#[derive(Debug, thiserror::Error)]
+#[non_exhaustive]
+pub enum StorageError {
+    /// Error from the OS I/O layer.
+    #[error(transparent)]
+    Io(#[from] io::Error),
+    /// Error related to data corruption.
+    #[error(transparent)]
+    Corruption(#[from] CorruptionError),
+    #[doc(hidden)]
+    #[error("{0}")]
+    Other(#[source] Box<dyn Error + Send + Sync + 'static>),
+}
+
+impl From<StorageError> for io::Error {
+    #[inline]
+    fn from(error: StorageError) -> Self {
+        match error {
+            StorageError::Io(error) => error,
+            StorageError::Corruption(error) => error.into(),
+            StorageError::Other(error) => Self::new(io::ErrorKind::Other, error),
+        }
+    }
+}
+
+/// An error return if some content in the database is corrupted.
+#[derive(Debug, thiserror::Error)]
+#[error(transparent)]
+pub struct CorruptionError(#[from] CorruptionErrorKind);
+
+/// An error return if some content in the database is corrupted.
+#[derive(Debug, thiserror::Error)]
+enum CorruptionErrorKind {
+    #[error("{0}")]
+    Msg(String),
+    #[error("{0}")]
+    Other(#[source] Box<dyn Error + Send + Sync + 'static>),
+}
+
+impl CorruptionError {
+    /// Builds an error from a printable error message.
+    #[inline]
+    pub(crate) fn new(error: impl Into<Box<dyn Error + Send + Sync + 'static>>) -> Self {
+        Self(CorruptionErrorKind::Other(error.into()))
+    }
+
+    #[inline]
+    pub(crate) fn from_encoded_term(encoded: &EncodedTerm, term: &TermRef<'_>) -> Self {
+        // TODO: eventually use a dedicated error enum value
+        Self::msg(format!("Invalid term encoding {encoded:?} for {term}"))
+    }
+
+    #[inline]
+    pub(crate) fn from_missing_column_family_name(name: &'static str) -> Self {
+        // TODO: eventually use a dedicated error enum value
+        Self::msg(format!("Column family {name} does not exist"))
+    }
+
+    /// Builds an error from a printable error message.
+    #[inline]
+    pub(crate) fn msg(msg: impl Into<String>) -> Self {
+        Self(CorruptionErrorKind::Msg(msg.into()))
+    }
+}
+
+impl From<CorruptionError> for io::Error {
+    #[inline]
+    fn from(error: CorruptionError) -> Self {
+        Self::new(io::ErrorKind::InvalidData, error)
+    }
+}
+
+/// An error raised while loading a file into a [`Store`](crate::store::Store).
+#[derive(Debug, thiserror::Error)]
+pub enum LoaderError {
+    /// An error raised while reading the file.
+    #[error(transparent)]
+    Parsing(#[from] RdfParseError),
+    /// An error raised during the insertion in the store.
+    #[error(transparent)]
+    Storage(#[from] StorageError),
+    /// The base IRI is invalid.
+    #[error("Invalid base IRI '{iri}': {error}")]
+    InvalidBaseIri {
+        /// The IRI itself.
+        iri: String,
+        /// The parsing error.
+        #[source]
+        error: IriParseError,
+    },
+}
+
+impl From<LoaderError> for io::Error {
+    #[inline]
+    fn from(error: LoaderError) -> Self {
+        match error {
+            LoaderError::Storage(error) => error.into(),
+            LoaderError::Parsing(error) => error.into(),
+            LoaderError::InvalidBaseIri { .. } => {
+                Self::new(io::ErrorKind::InvalidInput, error.to_string())
+            }
+        }
+    }
+}
+
+/// An error raised while writing a file from a [`Store`](crate::store::Store).
+#[derive(Debug, thiserror::Error)]
+pub enum SerializerError {
+    /// An error raised while writing the content.
+    #[error(transparent)]
+    Io(#[from] io::Error),
+    /// An error raised during the lookup in the store.
+    #[error(transparent)]
+    Storage(#[from] StorageError),
+    /// A format compatible with [RDF dataset](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-dataset) is required.
+    #[error("A RDF format supporting datasets was expected, {0} found")]
+    DatasetFormatExpected(RdfFormat),
+}
+
+impl From<SerializerError> for io::Error {
+    #[inline]
+    fn from(error: SerializerError) -> Self {
+        match error {
+            SerializerError::Storage(error) => error.into(),
+            SerializerError::Io(error) => error,
+            SerializerError::DatasetFormatExpected(_) => {
+                Self::new(io::ErrorKind::InvalidInput, error.to_string())
+            }
+        }
+    }
+}

lib/oxigraph/src/storage/mod.rs

@@ -1,9 +1,9 @@
 #![allow(clippy::same_name_method)]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use crate::model::Quad;
 use crate::model::{GraphNameRef, NamedOrBlankNodeRef, QuadRef, TermRef};
 use crate::storage::backend::{Reader, Transaction};
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use crate::storage::binary_encoder::LATEST_STORAGE_VERSION;
 use crate::storage::binary_encoder::{
     decode_term, encode_term, encode_term_pair, encode_term_quad, encode_term_triple,
@@ -12,27 +12,23 @@ use crate::storage::binary_encoder::{
     WRITTEN_TERM_MAX_SIZE,
 };
 pub use crate::storage::error::{CorruptionError, LoaderError, SerializerError, StorageError};
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use crate::storage::numeric_encoder::Decoder;
 use crate::storage::numeric_encoder::{insert_term, EncodedQuad, EncodedTerm, StrHash, StrLookup};
 use backend::{ColumnFamily, ColumnFamilyDefinition, Db, Iter};
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::collections::VecDeque;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::collections::{HashMap, HashSet};
 use std::error::Error;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::mem::{swap, take};
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::path::{Path, PathBuf};
-#[cfg(not(target_family = "wasm"))]
-use std::sync::atomic::{AtomicU64, Ordering};
-#[cfg(not(target_family = "wasm"))]
-use std::sync::Arc;
-#[cfg(not(target_family = "wasm"))]
-use std::thread::spawn;
-#[cfg(not(target_family = "wasm"))]
-use std::thread::JoinHandle;
+#[cfg(all(not(target_family = "wasm")))]
+use std::sync::Mutex;
+#[cfg(all(not(target_family = "wasm")))]
+use std::{io, thread};
 
 mod backend;
 mod binary_encoder;
@@ -51,16 +47,16 @@ const DSPO_CF: &str = "dspo";
 const DPOS_CF: &str = "dpos";
 const DOSP_CF: &str = "dosp";
 const GRAPHS_CF: &str = "graphs";
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 const DEFAULT_CF: &str = "default";
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 const DEFAULT_BULK_LOAD_BATCH_SIZE: usize = 1_000_000;
 
 /// Low level storage primitives
 #[derive(Clone)]
 pub struct Storage {
     db: Db,
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     default_cf: ColumnFamily,
     id2str_cf: ColumnFamily,
     spog_cf: ColumnFamily,
@@ -80,35 +76,39 @@ impl Storage {
         Self::setup(Db::new(Self::column_families())?)
     }
 
-    #[cfg(not(target_family = "wasm"))]
-    pub fn open(path: &Path) -> Result<Self, StorageError> {
-        Self::setup(Db::open_read_write(Some(path), Self::column_families())?)
-    }
-
-    #[cfg(not(target_family = "wasm"))]
-    pub fn open_secondary(primary_path: &Path) -> Result<Self, StorageError> {
-        Self::setup(Db::open_secondary(
-            primary_path,
-            None,
-            Self::column_families(),
-        )?)
-    }
-
-    #[cfg(not(target_family = "wasm"))]
-    pub fn open_persistent_secondary(
-        primary_path: &Path,
-        secondary_path: &Path,
-    ) -> Result<Self, StorageError> {
-        Self::setup(Db::open_secondary(
-            primary_path,
-            Some(secondary_path),
+    #[cfg(all(not(target_family = "wasm")))]
+    pub fn open(path: &Path, key: Option<[u8; 32]>) -> Result<Self, StorageError> {
+        Self::setup(Db::open_read_write(
+            Some(path),
             Self::column_families(),
+            key,
         )?)
     }
 
-    #[cfg(not(target_family = "wasm"))]
-    pub fn open_read_only(path: &Path) -> Result<Self, StorageError> {
-        Self::setup(Db::open_read_only(path, Self::column_families())?)
+    // #[cfg(all(not(target_family = "wasm")))]
+    // pub fn open_secondary(primary_path: &Path) -> Result<Self, StorageError> {
+    //     Self::setup(Db::open_secondary(
+    //         primary_path,
+    //         None,
+    //         Self::column_families(),
+    //     )?)
+    // }
+
+    // #[cfg(all(not(target_family = "wasm")))]
+    // pub fn open_persistent_secondary(
+    //     primary_path: &Path,
+    //     secondary_path: &Path,
+    // ) -> Result<Self, StorageError> {
+    //     Self::setup(Db::open_secondary(
+    //         primary_path,
+    //         Some(secondary_path),
+    //         Self::column_families(),
+    //     )?)
+    // }
+
+    #[cfg(all(not(target_family = "wasm")))]
+    pub fn open_read_only(path: &Path, key: Option<[u8; 32]>) -> Result<Self, StorageError> {
+        Self::setup(Db::open_read_only(path, Self::column_families(), key)?)
     }
 
     fn column_families() -> Vec<ColumnFamilyDefinition> {
@@ -182,30 +182,29 @@ impl Storage {
         ]
     }
 
-    #[allow(clippy::unnecessary_wraps, clippy::unwrap_in_result)]
     fn setup(db: Db) -> Result<Self, StorageError> {
         let this = Self {
-            #[cfg(not(target_family = "wasm"))]
-            default_cf: db.column_family(DEFAULT_CF).unwrap(),
-            id2str_cf: db.column_family(ID2STR_CF).unwrap(),
-            spog_cf: db.column_family(SPOG_CF).unwrap(),
-            posg_cf: db.column_family(POSG_CF).unwrap(),
-            ospg_cf: db.column_family(OSPG_CF).unwrap(),
-            gspo_cf: db.column_family(GSPO_CF).unwrap(),
-            gpos_cf: db.column_family(GPOS_CF).unwrap(),
-            gosp_cf: db.column_family(GOSP_CF).unwrap(),
-            dspo_cf: db.column_family(DSPO_CF).unwrap(),
-            dpos_cf: db.column_family(DPOS_CF).unwrap(),
-            dosp_cf: db.column_family(DOSP_CF).unwrap(),
-            graphs_cf: db.column_family(GRAPHS_CF).unwrap(),
+            #[cfg(all(not(target_family = "wasm")))]
+            default_cf: db.column_family(DEFAULT_CF)?,
+            id2str_cf: db.column_family(ID2STR_CF)?,
+            spog_cf: db.column_family(SPOG_CF)?,
+            posg_cf: db.column_family(POSG_CF)?,
+            ospg_cf: db.column_family(OSPG_CF)?,
+            gspo_cf: db.column_family(GSPO_CF)?,
+            gpos_cf: db.column_family(GPOS_CF)?,
+            gosp_cf: db.column_family(GOSP_CF)?,
+            dspo_cf: db.column_family(DSPO_CF)?,
+            dpos_cf: db.column_family(DPOS_CF)?,
+            dosp_cf: db.column_family(DOSP_CF)?,
+            graphs_cf: db.column_family(GRAPHS_CF)?,
             db,
         };
-        #[cfg(not(target_family = "wasm"))]
+        #[cfg(all(not(target_family = "wasm")))]
         this.migrate()?;
         Ok(this)
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     fn migrate(&self) -> Result<(), StorageError> {
         let mut version = self.ensure_version()?;
         if version == 0 {
@@ -245,7 +244,7 @@ impl Storage {
         }
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     fn ensure_version(&self) -> Result<u64, StorageError> {
         Ok(
             if let Some(version) = self.db.get(&self.default_cf, b"oxversion")? {
@@ -259,11 +258,11 @@ impl Storage {
         )
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     fn update_version(&self, version: u64) -> Result<(), StorageError> {
         self.db
             .insert(&self.default_cf, b"oxversion", &version.to_be_bytes())?;
-        self.db.flush(&self.default_cf)
+        self.db.flush()
     }
 
     pub fn snapshot(&self) -> StorageReader {
@@ -286,22 +285,12 @@ impl Storage {
         })
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn flush(&self) -> Result<(), StorageError> {
-        self.db.flush(&self.default_cf)?;
-        self.db.flush(&self.gspo_cf)?;
-        self.db.flush(&self.gpos_cf)?;
-        self.db.flush(&self.gosp_cf)?;
-        self.db.flush(&self.spog_cf)?;
-        self.db.flush(&self.posg_cf)?;
-        self.db.flush(&self.ospg_cf)?;
-        self.db.flush(&self.dspo_cf)?;
-        self.db.flush(&self.dpos_cf)?;
-        self.db.flush(&self.dosp_cf)?;
-        self.db.flush(&self.id2str_cf)
-    }
-
-    #[cfg(not(target_family = "wasm"))]
+        self.db.flush()
+    }
+
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn compact(&self) -> Result<(), StorageError> {
         self.db.compact(&self.default_cf)?;
         self.db.compact(&self.gspo_cf)?;
@@ -316,7 +305,7 @@ impl Storage {
         self.db.compact(&self.id2str_cf)
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn backup(&self, target_directory: &Path) -> Result<(), StorageError> {
         self.db.backup(target_directory)
     }
@@ -584,7 +573,7 @@ impl StorageReader {
 
     pub fn named_graphs(&self) -> DecodingGraphIterator {
         DecodingGraphIterator {
-            iter: self.reader.iter(&self.storage.graphs_cf).unwrap(), //TODO: propagate error?
+            iter: self.reader.iter(&self.storage.graphs_cf).unwrap(), // TODO: propagate error?
         }
     }
 
@@ -641,7 +630,7 @@ impl StorageReader {
         }
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn get_str(&self, key: &StrHash) -> Result<Option<String>, StorageError> {
         Ok(self
             .storage
@@ -652,7 +641,7 @@ impl StorageReader {
             .map_err(CorruptionError::new)?)
     }
 
-    #[cfg(target_family = "wasm")]
+    #[cfg(any(target_family = "wasm"))]
     pub fn get_str(&self, key: &StrHash) -> Result<Option<String>, StorageError> {
         Ok(self
             .reader
@@ -662,21 +651,21 @@ impl StorageReader {
             .map_err(CorruptionError::new)?)
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError> {
         self.storage
             .db
             .contains_key(&self.storage.id2str_cf, &key.to_be_bytes())
     }
 
-    #[cfg(target_family = "wasm")]
+    #[cfg(any(target_family = "wasm"))]
     pub fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError> {
         self.reader
             .contains_key(&self.storage.id2str_cf, &key.to_be_bytes())
     }
 
     /// Validates that all the storage invariants held in the data
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     pub fn validate(&self) -> Result<(), StorageError> {
         // triples
         let dspo_size = self.dspo_quads(&[]).count();
@@ -788,10 +777,10 @@ impl StorageReader {
     }
 
     /// Validates that all the storage invariants held in the data
-    #[cfg(target_family = "wasm")]
+    #[cfg(any(target_family = "wasm"))]
     #[allow(clippy::unused_self, clippy::unnecessary_wraps)]
     pub fn validate(&self) -> Result<(), StorageError> {
-        Ok(()) //TODO
+        Ok(()) // TODO
     }
 }
 
@@ -819,7 +808,7 @@ impl ChainedDecodingQuadIterator {
 impl Iterator for ChainedDecodingQuadIterator {
     type Item = Result<EncodedQuad, StorageError>;
 
-    fn next(&mut self) -> Option<Result<EncodedQuad, StorageError>> {
+    fn next(&mut self) -> Option<Self::Item> {
         if let Some(result) = self.first.next() {
             Some(result)
         } else if let Some(second) = self.second.as_mut() {
@@ -838,7 +827,7 @@ pub struct DecodingQuadIterator {
 impl Iterator for DecodingQuadIterator {
     type Item = Result<EncodedQuad, StorageError>;
 
-    fn next(&mut self) -> Option<Result<EncodedQuad, StorageError>> {
+    fn next(&mut self) -> Option<Self::Item> {
         if let Err(e) = self.iter.status() {
             return Some(Err(e));
         }
@@ -855,7 +844,7 @@ pub struct DecodingGraphIterator {
 impl Iterator for DecodingGraphIterator {
     type Item = Result<EncodedTerm, StorageError>;
 
-    fn next(&mut self) -> Option<Result<EncodedTerm, StorageError>> {
+    fn next(&mut self) -> Option<Self::Item> {
         if let Err(e) = self.iter.status() {
             return Some(Err(e));
         }
@@ -869,10 +858,6 @@ impl StrLookup for StorageReader {
     fn get_str(&self, key: &StrHash) -> Result<Option<String>, StorageError> {
         self.get_str(key)
     }
-
-    fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError> {
-        self.contains_str(key)
-    }
 }
 
 pub struct StorageWriter<'a> {
@@ -1016,7 +1001,7 @@ impl<'a> StorageWriter<'a> {
         }
     }
 
-    #[cfg(not(target_family = "wasm"))]
+    #[cfg(all(not(target_family = "wasm")))]
     fn insert_str(&mut self, key: &StrHash, value: &str) -> Result<(), StorageError> {
         if self
             .storage
@@ -1032,7 +1017,7 @@ impl<'a> StorageWriter<'a> {
         )
     }
 
-    #[cfg(target_family = "wasm")]
+    #[cfg(any(target_family = "wasm"))]
     fn insert_str(&mut self, key: &StrHash, value: &str) -> Result<(), StorageError> {
         self.transaction.insert(
             &self.storage.id2str_cf,
@@ -1197,7 +1182,8 @@ impl<'a> StorageWriter<'a> {
     }
 }
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
+#[must_use]
 pub struct StorageBulkLoader {
     storage: Storage,
     hooks: Vec<Box<dyn Fn(u64)>>,
@@ -1205,7 +1191,7 @@ pub struct StorageBulkLoader {
     max_memory_size: Option<usize>,
 }
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 impl StorageBulkLoader {
     pub fn new(storage: Storage) -> Self {
         Self {
@@ -1216,12 +1202,12 @@ impl StorageBulkLoader {
         }
     }
 
-    pub fn set_num_threads(mut self, num_threads: usize) -> Self {
+    pub fn with_num_threads(mut self, num_threads: usize) -> Self {
         self.num_threads = Some(num_threads);
         self
     }
 
-    pub fn set_max_memory_size_in_megabytes(mut self, max_memory_size: usize) -> Self {
+    pub fn with_max_memory_size_in_megabytes(mut self, max_memory_size: usize) -> Self {
         self.max_memory_size = Some(max_memory_size);
         self
     }
@@ -1253,90 +1239,101 @@ impl StorageBulkLoader {
             )
             .into());
         }
-        let mut threads = VecDeque::with_capacity(num_threads - 1);
-        let mut buffer = Vec::with_capacity(batch_size);
-        let done_counter = Arc::new(AtomicU64::new(0));
+        let done_counter = Mutex::new(0);
         let mut done_and_displayed_counter = 0;
-        for quad in quads {
-            let quad = quad?;
-            buffer.push(quad);
-            if buffer.len() >= batch_size {
-                self.spawn_load_thread(
-                    &mut buffer,
-                    &mut threads,
-                    &done_counter,
-                    &mut done_and_displayed_counter,
-                    num_threads,
-                    batch_size,
-                )?;
+        thread::scope(|thread_scope| {
+            let mut threads = VecDeque::with_capacity(num_threads - 1);
+            let mut buffer = Vec::with_capacity(batch_size);
+            for quad in quads {
+                let quad = quad?;
+                buffer.push(quad);
+                if buffer.len() >= batch_size {
+                    self.spawn_load_thread(
+                        &mut buffer,
+                        &mut threads,
+                        thread_scope,
+                        &done_counter,
+                        &mut done_and_displayed_counter,
+                        num_threads,
+                        batch_size,
+                    )?;
+                }
             }
-        }
-        self.spawn_load_thread(
-            &mut buffer,
-            &mut threads,
-            &done_counter,
-            &mut done_and_displayed_counter,
-            num_threads,
-            batch_size,
-        )?;
-        for thread in threads {
-            thread.join().unwrap()?;
-            self.on_possible_progress(&done_counter, &mut done_and_displayed_counter);
-        }
-        Ok(())
+            self.spawn_load_thread(
+                &mut buffer,
+                &mut threads,
+                thread_scope,
+                &done_counter,
+                &mut done_and_displayed_counter,
+                num_threads,
+                batch_size,
+            )?;
+            for thread in threads {
+                map_thread_result(thread.join()).map_err(StorageError::Io)??;
+                self.on_possible_progress(&done_counter, &mut done_and_displayed_counter)?;
+            }
+            Ok(())
+        })
     }
 
-    fn spawn_load_thread(
-        &self,
+    fn spawn_load_thread<'scope>(
+        &'scope self,
         buffer: &mut Vec<Quad>,
-        threads: &mut VecDeque<JoinHandle<Result<(), StorageError>>>,
-        done_counter: &Arc<AtomicU64>,
+        threads: &mut VecDeque<thread::ScopedJoinHandle<'scope, Result<(), StorageError>>>,
+        thread_scope: &'scope thread::Scope<'scope, '_>,
+        done_counter: &'scope Mutex<u64>,
         done_and_displayed_counter: &mut u64,
         num_threads: usize,
         batch_size: usize,
     ) -> Result<(), StorageError> {
-        self.on_possible_progress(done_counter, done_and_displayed_counter);
+        self.on_possible_progress(done_counter, done_and_displayed_counter)?;
         // We avoid to have too many threads
         if threads.len() >= num_threads {
             if let Some(thread) = threads.pop_front() {
-                thread.join().unwrap()?;
-                self.on_possible_progress(done_counter, done_and_displayed_counter);
+                map_thread_result(thread.join()).map_err(StorageError::Io)??;
+                self.on_possible_progress(done_counter, done_and_displayed_counter)?;
             }
         }
         let mut buffer_to_load = Vec::with_capacity(batch_size);
         swap(buffer, &mut buffer_to_load);
-        let storage = self.storage.clone();
-        let done_counter_clone = Arc::clone(done_counter);
-        threads.push_back(spawn(move || {
-            FileBulkLoader::new(storage, batch_size).load(buffer_to_load, &done_counter_clone)
+        let storage = &self.storage;
+        threads.push_back(thread_scope.spawn(move || {
+            FileBulkLoader::new(storage, batch_size).load(buffer_to_load, done_counter)
         }));
         Ok(())
     }
 
-    fn on_possible_progress(&self, done: &AtomicU64, done_and_displayed: &mut u64) {
-        let new_counter = done.load(Ordering::Relaxed);
-        let display_step = u64::try_from(DEFAULT_BULK_LOAD_BATCH_SIZE).unwrap();
+    fn on_possible_progress(
+        &self,
+        done: &Mutex<u64>,
+        done_and_displayed: &mut u64,
+    ) -> Result<(), StorageError> {
+        let new_counter = *done
+            .lock()
+            .map_err(|_| io::Error::new(io::ErrorKind::Other, "Mutex poisoned"))?;
+        let display_step = DEFAULT_BULK_LOAD_BATCH_SIZE as u64;
         if new_counter / display_step > *done_and_displayed / display_step {
             for hook in &self.hooks {
                 hook(new_counter);
             }
         }
         *done_and_displayed = new_counter;
+        Ok(())
     }
 }
 
-#[cfg(not(target_family = "wasm"))]
-struct FileBulkLoader {
-    storage: Storage,
+#[cfg(all(not(target_family = "wasm")))]
+struct FileBulkLoader<'a> {
+    storage: &'a Storage,
     id2str: HashMap<StrHash, Box<str>>,
     quads: HashSet<EncodedQuad>,
     triples: HashSet<EncodedQuad>,
     graphs: HashSet<EncodedTerm>,
 }
 
-#[cfg(not(target_family = "wasm"))]
-impl FileBulkLoader {
-    fn new(storage: Storage, batch_size: usize) -> Self {
+#[cfg(all(not(target_family = "wasm")))]
+impl<'a> FileBulkLoader<'a> {
+    fn new(storage: &'a Storage, batch_size: usize) -> Self {
         Self {
             storage,
             id2str: HashMap::with_capacity(3 * batch_size),
@@ -1346,11 +1343,14 @@ impl FileBulkLoader {
         }
     }
 
-    fn load(&mut self, quads: Vec<Quad>, counter: &AtomicU64) -> Result<(), StorageError> {
+    fn load(&mut self, quads: Vec<Quad>, counter: &Mutex<u64>) -> Result<(), StorageError> {
         self.encode(quads)?;
         let size = self.triples.len() + self.quads.len();
         self.save()?;
-        counter.fetch_add(size.try_into().unwrap(), Ordering::Relaxed);
+        *counter
+            .lock()
+            .map_err(|_| io::Error::new(io::ErrorKind::Other, "Mutex poisoned"))? +=
+            size.try_into().unwrap_or(u64::MAX);
         Ok(())
     }
 
@@ -1373,7 +1373,12 @@ impl FileBulkLoader {
                         match quad.graph_name.as_ref() {
                             GraphNameRef::NamedNode(n) => n.into(),
                             GraphNameRef::BlankNode(n) => n.into(),
-                            GraphNameRef::DefaultGraph => unreachable!(),
+                            GraphNameRef::DefaultGraph => {
+                                return Err(CorruptionError::new(
+                                    "Default graph this not the default graph",
+                                )
+                                .into())
+                            }
                         },
                         &encoded.graph_name,
                     )?;
@@ -1531,3 +1536,17 @@ impl FileBulkLoader {
         sst.finish()
     }
 }
+
+#[cfg(all(not(target_family = "wasm")))]
+fn map_thread_result<R>(result: thread::Result<R>) -> io::Result<R> {
+    result.map_err(|e| {
+        io::Error::new(
+            io::ErrorKind::Other,
+            if let Ok(e) = e.downcast::<&dyn std::fmt::Display>() {
+                format!("A loader processed crashed with {e}")
+            } else {
+                "A loader processed crashed with and unknown error".into()
+            },
+        )
+    })
+}

lib/oxigraph/src/storage/numeric_encoder.rs

@@ -6,10 +6,9 @@ use crate::storage::small_string::SmallString;
 use oxsdatatypes::*;
 use siphasher::sip128::{Hasher128, SipHasher24};
 use std::fmt::Debug;
-use std::hash::Hash;
-use std::hash::Hasher;
-use std::rc::Rc;
+use std::hash::{Hash, Hasher};
 use std::str;
+use std::sync::Arc;
 
 #[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
 #[repr(transparent)]
@@ -96,7 +95,7 @@ pub enum EncodedTerm {
     DurationLiteral(Duration),
     YearMonthDurationLiteral(YearMonthDuration),
     DayTimeDurationLiteral(DayTimeDuration),
-    Triple(Rc<EncodedTriple>),
+    Triple(Arc<EncodedTriple>),
 }
 
 impl PartialEq for EncodedTerm {
@@ -183,24 +182,24 @@ impl PartialEq for EncodedTerm {
                 },
             ) => value_id_a == value_id_b && datatype_id_a == datatype_id_b,
             (Self::BooleanLiteral(a), Self::BooleanLiteral(b)) => a == b,
-            (Self::FloatLiteral(a), Self::FloatLiteral(b)) => a.is_identical_with(b),
-            (Self::DoubleLiteral(a), Self::DoubleLiteral(b)) => a.is_identical_with(b),
-            (Self::IntegerLiteral(a), Self::IntegerLiteral(b)) => a.is_identical_with(b),
-            (Self::DecimalLiteral(a), Self::DecimalLiteral(b)) => a.is_identical_with(b),
-            (Self::DateTimeLiteral(a), Self::DateTimeLiteral(b)) => a.is_identical_with(b),
-            (Self::TimeLiteral(a), Self::TimeLiteral(b)) => a.is_identical_with(b),
-            (Self::DateLiteral(a), Self::DateLiteral(b)) => a.is_identical_with(b),
-            (Self::GYearMonthLiteral(a), Self::GYearMonthLiteral(b)) => a.is_identical_with(b),
-            (Self::GYearLiteral(a), Self::GYearLiteral(b)) => a.is_identical_with(b),
-            (Self::GMonthDayLiteral(a), Self::GMonthDayLiteral(b)) => a.is_identical_with(b),
-            (Self::GMonthLiteral(a), Self::GMonthLiteral(b)) => a.is_identical_with(b),
-            (Self::GDayLiteral(a), Self::GDayLiteral(b)) => a.is_identical_with(b),
-            (Self::DurationLiteral(a), Self::DurationLiteral(b)) => a.is_identical_with(b),
+            (Self::FloatLiteral(a), Self::FloatLiteral(b)) => a.is_identical_with(*b),
+            (Self::DoubleLiteral(a), Self::DoubleLiteral(b)) => a.is_identical_with(*b),
+            (Self::IntegerLiteral(a), Self::IntegerLiteral(b)) => a.is_identical_with(*b),
+            (Self::DecimalLiteral(a), Self::DecimalLiteral(b)) => a.is_identical_with(*b),
+            (Self::DateTimeLiteral(a), Self::DateTimeLiteral(b)) => a.is_identical_with(*b),
+            (Self::TimeLiteral(a), Self::TimeLiteral(b)) => a.is_identical_with(*b),
+            (Self::DateLiteral(a), Self::DateLiteral(b)) => a.is_identical_with(*b),
+            (Self::GYearMonthLiteral(a), Self::GYearMonthLiteral(b)) => a.is_identical_with(*b),
+            (Self::GYearLiteral(a), Self::GYearLiteral(b)) => a.is_identical_with(*b),
+            (Self::GMonthDayLiteral(a), Self::GMonthDayLiteral(b)) => a.is_identical_with(*b),
+            (Self::GMonthLiteral(a), Self::GMonthLiteral(b)) => a.is_identical_with(*b),
+            (Self::GDayLiteral(a), Self::GDayLiteral(b)) => a.is_identical_with(*b),
+            (Self::DurationLiteral(a), Self::DurationLiteral(b)) => a.is_identical_with(*b),
             (Self::YearMonthDurationLiteral(a), Self::YearMonthDurationLiteral(b)) => {
-                a.is_identical_with(b)
+                a.is_identical_with(*b)
             }
             (Self::DayTimeDurationLiteral(a), Self::DayTimeDurationLiteral(b)) => {
-                a.is_identical_with(b)
+                a.is_identical_with(*b)
             }
             (Self::Triple(a), Self::Triple(b)) => a == b,
             (_, _) => false,
@@ -471,7 +470,7 @@ impl From<DayTimeDuration> for EncodedTerm {
 
 impl From<EncodedTriple> for EncodedTerm {
     fn from(value: EncodedTriple) -> Self {
-        Self::Triple(Rc::new(value))
+        Self::Triple(Arc::new(value))
     }
 }
 
@@ -634,7 +633,7 @@ impl From<GraphNameRef<'_>> for EncodedTerm {
 
 impl From<TripleRef<'_>> for EncodedTerm {
     fn from(triple: TripleRef<'_>) -> Self {
-        Self::Triple(Rc::new(triple.into()))
+        Self::Triple(Arc::new(triple.into()))
     }
 }
 
@@ -702,8 +701,6 @@ impl From<QuadRef<'_>> for EncodedQuad {
 
 pub trait StrLookup {
     fn get_str(&self, key: &StrHash) -> Result<Option<String>, StorageError>;
-
-    fn contains_str(&self, key: &StrHash) -> Result<bool, StorageError>;
 }
 
 pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
@@ -716,13 +713,13 @@ pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
             if let EncodedTerm::NamedNode { iri_id } = encoded {
                 insert_str(iri_id, node.as_str())
             } else {
-                unreachable!("Invalid term encoding {:?} for {}", encoded, term)
+                Err(CorruptionError::from_encoded_term(encoded, &term).into())
             }
         }
         TermRef::BlankNode(node) => match encoded {
             EncodedTerm::BigBlankNode { id_id } => insert_str(id_id, node.as_str()),
             EncodedTerm::SmallBlankNode(..) | EncodedTerm::NumericalBlankNode { .. } => Ok(()),
-            _ => unreachable!("Invalid term encoding {:?} for {}", encoded, term),
+            _ => Err(CorruptionError::from_encoded_term(encoded, &term).into()),
         },
         TermRef::Literal(literal) => match encoded {
             EncodedTerm::BigStringLiteral { value_id }
@@ -733,7 +730,7 @@ pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
                 if let Some(language) = literal.language() {
                     insert_str(language_id, language)
                 } else {
-                    unreachable!("Invalid term encoding {:?} for {}", encoded, term)
+                    Err(CorruptionError::from_encoded_term(encoded, &term).into())
                 }
             }
             EncodedTerm::BigBigLangStringLiteral {
@@ -744,7 +741,7 @@ pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
                 if let Some(language) = literal.language() {
                     insert_str(language_id, language)
                 } else {
-                    unreachable!("Invalid term encoding {:?} for {}", encoded, term)
+                    Err(CorruptionError::from_encoded_term(encoded, &term).into())
                 }
             }
             EncodedTerm::SmallTypedLiteral { datatype_id, .. } => {
@@ -775,7 +772,7 @@ pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
             | EncodedTerm::DurationLiteral(..)
             | EncodedTerm::YearMonthDurationLiteral(..)
             | EncodedTerm::DayTimeDurationLiteral(..) => Ok(()),
-            _ => unreachable!("Invalid term encoding {:?} for {}", encoded, term),
+            _ => Err(CorruptionError::from_encoded_term(encoded, &term).into()),
         },
         TermRef::Triple(triple) => {
             if let EncodedTerm::Triple(encoded) = encoded {
@@ -787,7 +784,7 @@ pub fn insert_term<F: FnMut(&StrHash, &str) -> Result<(), StorageError>>(
                 )?;
                 insert_term(triple.object.as_ref(), &encoded.object, insert_str)
             } else {
-                unreachable!("Invalid term encoding {:?} for {}", encoded, term)
+                Err(CorruptionError::from_encoded_term(encoded, &term).into())
             }
         }
     }

lib/oxigraph/src/storage/small_string.rs

@@ -1,11 +1,9 @@
 use std::borrow::Borrow;
 use std::cmp::Ordering;
-use std::error::Error;
-use std::fmt;
 use std::hash::{Hash, Hasher};
 use std::ops::Deref;
-use std::str;
 use std::str::{FromStr, Utf8Error};
+use std::{fmt, str};
 
 /// A small inline string
 #[derive(Clone, Copy, Default)]
@@ -46,10 +44,8 @@ impl SmallString {
     #[inline]
     #[allow(unsafe_code)]
     pub fn as_str(&self) -> &str {
-        unsafe {
-            // safe because we ensured it in constructors
-            str::from_utf8_unchecked(self.as_bytes())
-        }
+        // SAFETY: safe because we ensured it in constructors
+        unsafe { str::from_utf8_unchecked(self.as_bytes()) }
     }
 
     #[inline]
@@ -67,7 +63,7 @@ impl Deref for SmallString {
     type Target = str;
 
     #[inline]
-    fn deref(&self) -> &str {
+    fn deref(&self) -> &Self::Target {
         self.as_str()
     }
 }
@@ -103,7 +99,7 @@ impl fmt::Display for SmallString {
 impl PartialEq for SmallString {
     #[inline]
     fn eq(&self, other: &Self) -> bool {
-        self.as_str().eq(&**other)
+        self.as_str() == other.as_str()
     }
 }
 
@@ -148,17 +144,17 @@ impl FromStr for SmallString {
     type Err = BadSmallStringError;
 
     #[inline]
-    fn from_str(value: &str) -> Result<Self, BadSmallStringError> {
+    fn from_str(value: &str) -> Result<Self, Self::Err> {
         if value.len() <= 15 {
             let mut inner = [0; 16];
             inner[..value.len()].copy_from_slice(value.as_bytes());
             inner[15] = value
                 .len()
                 .try_into()
-                .map_err(|_| BadSmallStringError::TooLong(value.len()))?;
+                .map_err(|_| Self::Err::TooLong(value.len()))?;
             Ok(Self { inner })
         } else {
-            Err(BadSmallStringError::TooLong(value.len()))
+            Err(Self::Err::TooLong(value.len()))
         }
     }
 }
@@ -167,36 +163,15 @@ impl<'a> TryFrom<&'a str> for SmallString {
     type Error = BadSmallStringError;
 
     #[inline]
-    fn try_from(value: &'a str) -> Result<Self, BadSmallStringError> {
+    fn try_from(value: &'a str) -> Result<Self, Self::Error> {
         Self::from_str(value)
     }
 }
 
-#[derive(Debug, Clone, Copy)]
+#[derive(Debug, Clone, Copy, thiserror::Error)]
 pub enum BadSmallStringError {
+    #[error("small strings could only contain at most 15 characters, found {0}")]
     TooLong(usize),
-    BadUtf8(Utf8Error),
-}
-
-impl fmt::Display for BadSmallStringError {
-    #[inline]
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        match self {
-            Self::TooLong(v) => write!(
-                f,
-                "small strings could only contain at most 15 characters, found {v}"
-            ),
-            Self::BadUtf8(e) => e.fmt(f),
-        }
-    }
-}
-
-impl Error for BadSmallStringError {
-    #[inline]
-    fn source(&self) -> Option<&(dyn Error + 'static)> {
-        match self {
-            Self::TooLong(_) => None,
-            Self::BadUtf8(e) => Some(e),
-        }
-    }
+    #[error(transparent)]
+    BadUtf8(#[from] Utf8Error),
 }

lib/oxigraph/tests/store.rs

@@ -1,24 +1,26 @@
-use oxigraph::io::{DatasetFormat, GraphFormat};
+#![cfg(test)]
+#![allow(clippy::panic_in_result_fn)]
+
+use oxigraph::io::RdfFormat;
 use oxigraph::model::vocab::{rdf, xsd};
 use oxigraph::model::*;
 use oxigraph::store::Store;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use rand::random;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::env::temp_dir;
 use std::error::Error;
-#[cfg(not(target_family = "wasm"))]
-use std::fs::{create_dir, remove_dir_all, File};
-use std::io::Cursor;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
+use std::fs::{create_dir_all, remove_dir_all, File};
+#[cfg(all(not(target_family = "wasm")))]
 use std::io::Write;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::iter::empty;
-#[cfg(target_os = "linux")]
+#[cfg(all(target_os = "linux"))]
 use std::iter::once;
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 use std::path::{Path, PathBuf};
-#[cfg(target_os = "linux")]
+#[cfg(all(target_os = "linux"))]
 use std::process::Command;
 
 #[allow(clippy::non_ascii_literal)]
@@ -76,7 +78,7 @@ fn quads(graph_name: impl Into<GraphNameRef<'static>>) -> Vec<QuadRef<'static>>
         QuadRef::new(
             paris,
             name,
-            LiteralRef::new_language_tagged_literal_unchecked("la ville lumi\u{e8}re", "fr"),
+            LiteralRef::new_language_tagged_literal_unchecked("la ville lumi\u{E8}re", "fr"),
             graph_name,
         ),
         QuadRef::new(paris, country, france, graph_name),
@@ -110,12 +112,7 @@ fn quads(graph_name: impl Into<GraphNameRef<'static>>) -> Vec<QuadRef<'static>>
 #[test]
 fn test_load_graph() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
-    store.load_graph(
-        Cursor::new(DATA),
-        GraphFormat::Turtle,
-        GraphNameRef::DefaultGraph,
-        None,
-    )?;
+    store.load_from_read(RdfFormat::Turtle, DATA.as_bytes())?;
     for q in quads(GraphNameRef::DefaultGraph) {
         assert!(store.contains(q)?);
     }
@@ -124,15 +121,12 @@ fn test_load_graph() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bulk_load_graph() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
-    store.bulk_loader().load_graph(
-        Cursor::new(DATA),
-        GraphFormat::Turtle,
-        GraphNameRef::DefaultGraph,
-        None,
-    )?;
+    store
+        .bulk_loader()
+        .load_from_read(RdfFormat::Turtle, DATA.as_bytes())?;
     for q in quads(GraphNameRef::DefaultGraph) {
         assert!(store.contains(q)?);
     }
@@ -141,14 +135,12 @@ fn test_bulk_load_graph() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bulk_load_graph_lenient() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
-    store.bulk_loader().on_parse_error(|_| Ok(())).load_graph(
-        Cursor::new(b"<http://example.com> <http://example.com> <http://example.com##> .\n<http://example.com> <http://example.com> <http://example.com> ."),
-        GraphFormat::NTriples,
-        GraphNameRef::DefaultGraph,
-        None,
+    store.bulk_loader().on_parse_error(|_| Ok(())).load_from_read(
+        RdfFormat::NTriples,
+        b"<http://example.com> <http://example.com> <http://example.com##> .\n<http://example.com> <http://example.com> <http://example.com> .".as_slice(),
     )?;
     assert_eq!(store.len()?, 1);
     assert!(store.contains(QuadRef::new(
@@ -162,7 +154,7 @@ fn test_bulk_load_graph_lenient() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bulk_load_empty() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
     store.bulk_loader().load_quads(empty::<Quad>())?;
@@ -174,7 +166,7 @@ fn test_bulk_load_empty() -> Result<(), Box<dyn Error>> {
 #[test]
 fn test_load_dataset() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
-    store.load_dataset(Cursor::new(GRAPH_DATA), DatasetFormat::TriG, None)?;
+    store.load_from_read(RdfFormat::TriG, GRAPH_DATA.as_bytes())?;
     for q in quads(NamedNodeRef::new_unchecked(
         "http://www.wikidata.org/wiki/Special:EntityData/Q90",
     )) {
@@ -185,12 +177,12 @@ fn test_load_dataset() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bulk_load_dataset() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
     store
         .bulk_loader()
-        .load_dataset(Cursor::new(GRAPH_DATA), DatasetFormat::TriG, None)?;
+        .load_from_read(RdfFormat::TriG, GRAPH_DATA.as_bytes())?;
     let graph_name =
         NamedNodeRef::new_unchecked("http://www.wikidata.org/wiki/Special:EntityData/Q90");
     for q in quads(graph_name) {
@@ -205,11 +197,9 @@ fn test_bulk_load_dataset() -> Result<(), Box<dyn Error>> {
 fn test_load_graph_generates_new_blank_nodes() -> Result<(), Box<dyn Error>> {
     let store = Store::new()?;
     for _ in 0..2 {
-        store.load_graph(
-            Cursor::new("_:a <http://example.com/p> <http://example.com/p> ."),
-            GraphFormat::NTriples,
-            GraphNameRef::DefaultGraph,
-            None,
+        store.load_from_read(
+            RdfFormat::NTriples,
+            "_:a <http://example.com/p> <http://example.com/p> .".as_bytes(),
         )?;
     }
     assert_eq!(store.len()?, 2);
@@ -224,11 +214,7 @@ fn test_dump_graph() -> Result<(), Box<dyn Error>> {
     }
 
     let mut buffer = Vec::new();
-    store.dump_graph(
-        &mut buffer,
-        GraphFormat::NTriples,
-        GraphNameRef::DefaultGraph,
-    )?;
+    store.dump_graph_to_write(GraphNameRef::DefaultGraph, RdfFormat::NTriples, &mut buffer)?;
     assert_eq!(
         buffer.into_iter().filter(|c| *c == b'\n').count(),
         NUMBER_OF_TRIPLES
@@ -243,8 +229,7 @@ fn test_dump_dataset() -> Result<(), Box<dyn Error>> {
         store.insert(q)?;
     }
 
-    let mut buffer = Vec::new();
-    store.dump_dataset(&mut buffer, DatasetFormat::NQuads)?;
+    let buffer = store.dump_to_write(RdfFormat::NQuads, Vec::new())?;
     assert_eq!(
         buffer.into_iter().filter(|c| *c == b'\n').count(),
         NUMBER_OF_TRIPLES
@@ -255,10 +240,10 @@ fn test_dump_dataset() -> Result<(), Box<dyn Error>> {
 #[test]
 fn test_snapshot_isolation_iterator() -> Result<(), Box<dyn Error>> {
     let quad = QuadRef::new(
-        NamedNodeRef::new_unchecked("http://example.com/s"),
-        NamedNodeRef::new_unchecked("http://example.com/p"),
-        NamedNodeRef::new_unchecked("http://example.com/o"),
-        NamedNodeRef::new_unchecked("http://www.wikidata.org/wiki/Special:EntityData/Q90"),
+        NamedNodeRef::new("http://example.com/s")?,
+        NamedNodeRef::new("http://example.com/p")?,
+        NamedNodeRef::new("http://example.com/o")?,
+        NamedNodeRef::new("http://www.wikidata.org/wiki/Special:EntityData/Q90")?,
     );
     let store = Store::new()?;
     store.insert(quad)?;
@@ -273,7 +258,7 @@ fn test_snapshot_isolation_iterator() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bulk_load_on_existing_delete_overrides_the_delete() -> Result<(), Box<dyn Error>> {
     let quad = QuadRef::new(
         NamedNodeRef::new_unchecked("http://example.com/s"),
@@ -289,10 +274,10 @@ fn test_bulk_load_on_existing_delete_overrides_the_delete() -> Result<(), Box<dy
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_open_bad_dir() -> Result<(), Box<dyn Error>> {
     let dir = TempDir::default();
-    create_dir(&dir.0)?;
+    create_dir_all(&dir.0)?;
     {
         File::create(dir.0.join("CURRENT"))?.write_all(b"foo")?;
     }
@@ -301,84 +286,84 @@ fn test_open_bad_dir() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(target_os = "linux")]
+#[cfg(all(target_os = "linux"))]
 fn test_bad_stt_open() -> Result<(), Box<dyn Error>> {
     let dir = TempDir::default();
     let store = Store::open(&dir.0)?;
     remove_dir_all(&dir.0)?;
-    assert!(store
+    store
         .bulk_loader()
         .load_quads(once(Quad::new(
             NamedNode::new_unchecked("http://example.com/s"),
             NamedNode::new_unchecked("http://example.com/p"),
             NamedNode::new_unchecked("http://example.com/o"),
-            GraphName::DefaultGraph
+            GraphName::DefaultGraph,
         )))
-        .is_err());
+        .unwrap_err();
     Ok(())
 }
 
-#[test]
-#[cfg(not(target_family = "wasm"))]
-fn test_backup() -> Result<(), Box<dyn Error>> {
-    let quad = QuadRef::new(
-        NamedNodeRef::new_unchecked("http://example.com/s"),
-        NamedNodeRef::new_unchecked("http://example.com/p"),
-        NamedNodeRef::new_unchecked("http://example.com/o"),
-        GraphNameRef::DefaultGraph,
-    );
-    let store_dir = TempDir::default();
-    let backup_from_rw_dir = TempDir::default();
-    let backup_from_ro_dir = TempDir::default();
-    let backup_from_secondary_dir = TempDir::default();
-
-    let store = Store::open(&store_dir)?;
-    store.insert(quad)?;
-    let secondary_store = Store::open_secondary(&store_dir)?;
-    store.flush()?;
-
-    store.backup(&backup_from_rw_dir)?;
-    secondary_store.backup(&backup_from_secondary_dir)?;
-    store.remove(quad)?;
-    assert!(!store.contains(quad)?);
-
-    let backup_from_rw = Store::open_read_only(&backup_from_rw_dir.0)?;
-    backup_from_rw.validate()?;
-    assert!(backup_from_rw.contains(quad)?);
-    backup_from_rw.backup(&backup_from_ro_dir)?;
-
-    let backup_from_ro = Store::open_read_only(&backup_from_ro_dir.0)?;
-    backup_from_ro.validate()?;
-    assert!(backup_from_ro.contains(quad)?);
-
-    let backup_from_secondary = Store::open_read_only(&backup_from_secondary_dir.0)?;
-    backup_from_secondary.validate()?;
-    assert!(backup_from_secondary.contains(quad)?);
-
-    Ok(())
-}
+// #[test]
+// #[cfg(all(not(target_family = "wasm")))]
+// fn test_backup() -> Result<(), Box<dyn Error>> {
+//     let quad = QuadRef::new(
+//         NamedNodeRef::new_unchecked("http://example.com/s"),
+//         NamedNodeRef::new_unchecked("http://example.com/p"),
+//         NamedNodeRef::new_unchecked("http://example.com/o"),
+//         GraphNameRef::DefaultGraph,
+//     );
+//     let store_dir = TempDir::default();
+//     let backup_from_rw_dir = TempDir::default();
+//     let backup_from_ro_dir = TempDir::default();
+//     let backup_from_secondary_dir = TempDir::default();
+
+//     let store = Store::open(&store_dir)?;
+//     store.insert(quad)?;
+//     let secondary_store = Store::open_secondary(&store_dir)?;
+//     store.flush()?;
+
+//     store.backup(&backup_from_rw_dir)?;
+//     secondary_store.backup(&backup_from_secondary_dir)?;
+//     store.remove(quad)?;
+//     assert!(!store.contains(quad)?);
+
+//     let backup_from_rw = Store::open_read_only(&backup_from_rw_dir.0)?;
+//     backup_from_rw.validate()?;
+//     assert!(backup_from_rw.contains(quad)?);
+//     backup_from_rw.backup(&backup_from_ro_dir)?;
+
+//     let backup_from_ro = Store::open_read_only(&backup_from_ro_dir.0)?;
+//     backup_from_ro.validate()?;
+//     assert!(backup_from_ro.contains(quad)?);
+
+//     let backup_from_secondary = Store::open_read_only(&backup_from_secondary_dir.0)?;
+//     backup_from_secondary.validate()?;
+//     assert!(backup_from_secondary.contains(quad)?);
+
+//     Ok(())
+// }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_bad_backup() -> Result<(), Box<dyn Error>> {
     let store_dir = TempDir::default();
     let backup_dir = TempDir::default();
 
-    create_dir(&backup_dir.0)?;
-    assert!(Store::open(&store_dir)?.backup(&backup_dir.0).is_err());
+    create_dir_all(&backup_dir.0)?;
+    Store::open(&store_dir)?.backup(&backup_dir.0).unwrap_err();
     Ok(())
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_backup_on_in_memory() -> Result<(), Box<dyn Error>> {
     let backup_dir = TempDir::default();
-    assert!(Store::new()?.backup(&backup_dir).is_err());
+    Store::new()?.backup(&backup_dir).unwrap_err();
     Ok(())
 }
 
 #[test]
-#[cfg(target_os = "linux")]
+#[cfg(all(target_os = "linux"))]
 fn test_backward_compatibility() -> Result<(), Box<dyn Error>> {
     // We run twice to check if data is properly saved and closed
     for _ in 0..2 {
@@ -401,63 +386,63 @@ fn test_backward_compatibility() -> Result<(), Box<dyn Error>> {
     Ok(())
 }
 
-#[test]
-#[cfg(not(target_family = "wasm"))]
-fn test_secondary() -> Result<(), Box<dyn Error>> {
-    let quad = QuadRef::new(
-        NamedNodeRef::new_unchecked("http://example.com/s"),
-        NamedNodeRef::new_unchecked("http://example.com/p"),
-        NamedNodeRef::new_unchecked("http://example.com/o"),
-        GraphNameRef::DefaultGraph,
-    );
-    let primary_dir = TempDir::default();
-
-    // We open the store
-    let primary = Store::open(&primary_dir)?;
-    let secondary = Store::open_secondary(&primary_dir)?;
-
-    // We insert a quad
-    primary.insert(quad)?;
-    primary.flush()?;
-
-    // It is readable from both stores
-    for store in &[&primary, &secondary] {
-        assert!(store.contains(quad)?);
-        assert_eq!(
-            store.iter().collect::<Result<Vec<_>, _>>()?,
-            vec![quad.into_owned()]
-        );
-    }
-
-    // We validate the states
-    primary.validate()?;
-    secondary.validate()?;
-
-    // We close the primary store and remove its content
-    drop(primary);
-    remove_dir_all(&primary_dir)?;
-
-    // We secondary store is still readable
-    assert!(secondary.contains(quad)?);
-    secondary.validate()?;
-
-    Ok(())
-}
-
-#[test]
-#[cfg(not(target_family = "wasm"))]
-fn test_open_secondary_bad_dir() -> Result<(), Box<dyn Error>> {
-    let primary_dir = TempDir::default();
-    create_dir(&primary_dir.0)?;
-    {
-        File::create(primary_dir.0.join("CURRENT"))?.write_all(b"foo")?;
-    }
-    assert!(Store::open_secondary(&primary_dir).is_err());
-    Ok(())
-}
+// #[test]
+// #[cfg(all(not(target_family = "wasm")))]
+// fn test_secondary() -> Result<(), Box<dyn Error>> {
+//     let quad = QuadRef::new(
+//         NamedNodeRef::new_unchecked("http://example.com/s"),
+//         NamedNodeRef::new_unchecked("http://example.com/p"),
+//         NamedNodeRef::new_unchecked("http://example.com/o"),
+//         GraphNameRef::DefaultGraph,
+//     );
+//     let primary_dir = TempDir::default();
+
+//     // We open the store
+//     let primary = Store::open(&primary_dir)?;
+//     let secondary = Store::open_secondary(&primary_dir)?;
+
+//     // We insert a quad
+//     primary.insert(quad)?;
+//     primary.flush()?;
+
+//     // It is readable from both stores
+//     for store in &[&primary, &secondary] {
+//         assert!(store.contains(quad)?);
+//         assert_eq!(
+//             store.iter().collect::<Result<Vec<_>, _>>()?,
+//             vec![quad.into_owned()]
+//         );
+//     }
+
+//     // We validate the states
+//     primary.validate()?;
+//     secondary.validate()?;
+
+//     // We close the primary store and remove its content
+//     drop(primary);
+//     remove_dir_all(&primary_dir)?;
+
+//     // We secondary store is still readable
+//     assert!(secondary.contains(quad)?);
+//     secondary.validate()?;
+
+//     Ok(())
+// }
+
+// #[test]
+// #[cfg(all(not(target_family = "wasm")))]
+// fn test_open_secondary_bad_dir() -> Result<(), Box<dyn Error>> {
+//     let primary_dir = TempDir::default();
+//     create_dir_all(&primary_dir.0)?;
+//     {
+//         File::create(primary_dir.0.join("CURRENT"))?.write_all(b"foo")?;
+//     }
+//     assert!(Store::open_secondary(&primary_dir).is_err());
+//     Ok(())
+// }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_read_only() -> Result<(), Box<dyn Error>> {
     let s = NamedNodeRef::new_unchecked("http://example.com/s");
     let p = NamedNodeRef::new_unchecked("http://example.com/p");
@@ -483,7 +468,7 @@ fn test_read_only() -> Result<(), Box<dyn Error>> {
     }
 
     // We open as read-only
-    let read_only = Store::open_read_only(&store_dir)?;
+    let read_only = Store::open_read_only(&store_dir, None)?;
     assert!(read_only.contains(first_quad)?);
     assert_eq!(
         read_only.iter().collect::<Result<Vec<_>, _>>()?,
@@ -506,18 +491,18 @@ fn test_read_only() -> Result<(), Box<dyn Error>> {
 }
 
 #[test]
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 fn test_open_read_only_bad_dir() -> Result<(), Box<dyn Error>> {
     let dir = TempDir::default();
-    create_dir(&dir.0)?;
+    create_dir_all(&dir.0)?;
     {
         File::create(dir.0.join("CURRENT"))?.write_all(b"foo")?;
     }
-    assert!(Store::open_read_only(&dir).is_err());
+    assert!(Store::open_read_only(&dir, None).is_err());
     Ok(())
 }
 
-#[cfg(target_os = "linux")]
+#[cfg(all(target_os = "linux"))]
 fn reset_dir(dir: &str) -> Result<(), Box<dyn Error>> {
     assert!(Command::new("git")
         .args(["clean", "-fX", dir])
@@ -530,24 +515,24 @@ fn reset_dir(dir: &str) -> Result<(), Box<dyn Error>> {
     Ok(())
 }
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 struct TempDir(PathBuf);
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 impl Default for TempDir {
     fn default() -> Self {
         Self(temp_dir().join(format!("oxigraph-test-{}", random::<u128>())))
     }
 }
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 impl AsRef<Path> for TempDir {
     fn as_ref(&self) -> &Path {
         &self.0
     }
 }
 
-#[cfg(not(target_family = "wasm"))]
+#[cfg(all(not(target_family = "wasm")))]
 impl Drop for TempDir {
     fn drop(&mut self) {
         if self.0.is_dir() {

lib/oxrdf/Cargo.toml

@@ -1,28 +1,33 @@
 [package]
 name = "oxrdf"
-version = "0.1.7"
-authors = ["Tpt <thomas@pellissier-tanon.fr>"]
-license = "MIT OR Apache-2.0"
+version = "0.2.0-alpha.4"
+authors.workspace = true
+license.workspace = true
 readme = "README.md"
 keywords = ["RDF"]
 repository = "https://github.com/oxigraph/oxigraph/tree/main/lib/oxrdf"
-homepage = "https://oxigraph.org/"
 description = """
 A library providing basic data structures related to RDF
 """
 documentation = "https://docs.rs/oxrdf"
-edition = "2021"
-rust-version = "1.60"
+edition.workspace = true
+rust-version.workspace = true
 
 [features]
 default = []
 rdf-star = []
 
 [dependencies]
-rand = "0.8"
-oxilangtag = "0.1"
-oxiri = "0.2"
-oxsdatatypes = { version = "0.1.3", path="../oxsdatatypes", optional = true }
+oxilangtag.workspace = true
+oxiri.workspace = true
+oxsdatatypes = { workspace = true, optional = true }
+rand.workspace = true
+thiserror.workspace = true
+serde.workspace = true
+
+[lints]
+workspace = true
 
 [package.metadata.docs.rs]
 all-features = true
+rustdoc-args = ["--cfg", "docsrs"]

lib/oxrdf/README.md

@@ -5,7 +5,7 @@ OxRDF
 [![Released API docs](https://docs.rs/oxrdf/badge.svg)](https://docs.rs/oxrdf)
 [![Crates.io downloads](https://img.shields.io/crates/d/oxrdf)](https://crates.io/crates/oxrdf)
 [![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
-[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge)
+[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community)
 
 OxRDF is a simple library providing datastructures encoding [RDF 1.1 concepts](https://www.w3.org/TR/rdf11-concepts/).
 
@@ -15,6 +15,8 @@ Support for [RDF-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html) i
 
 OxRDF is inspired by [RDF/JS](https://rdf.js.org/data-model-spec/) and [Apache Commons RDF](http://commons.apache.org/proper/commons-rdf/).
 
+Use [`oxrdfio`](https://crates.io/crates/oxrdfio) if you need to read or write RDF files.
+
 Usage example:
 
 ```rust

lib/oxrdf/src/blank_node.rs

@@ -1,8 +1,7 @@
 use rand::random;
-use std::error::Error;
-use std::fmt;
+use serde::{Deserialize, Serialize};
 use std::io::Write;
-use std::str;
+use std::{fmt, str};
 
 /// An owned RDF [blank node](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node).
 ///
@@ -15,16 +14,13 @@ use std::str;
 /// ```
 /// use oxrdf::BlankNode;
 ///
-/// assert_eq!(
-///     "_:a122",
-///     BlankNode::new("a122")?.to_string()
-/// );
+/// assert_eq!("_:a122", BlankNode::new("a122")?.to_string());
 /// # Result::<_,oxrdf::BlankNodeIdParseError>::Ok(())
 /// ```
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub struct BlankNode(BlankNodeContent);
 
-#[derive(PartialEq, Eq, Debug, Clone, Hash)]
+#[derive(PartialEq, Eq, Debug, Clone, Hash, Serialize, Deserialize)]
 enum BlankNodeContent {
     Named(String),
     Anonymous { id: u128, str: IdStr },
@@ -36,7 +32,7 @@ impl BlankNode {
     /// The blank node identifier must be valid according to N-Triples, Turtle, and SPARQL grammars.
     ///
     /// In most cases, it is much more convenient to create a blank node using [`BlankNode::default()`]
-    ///that creates a random ID that could be easily inlined by Oxigraph stores.
+    /// that creates a random ID that could be easily inlined by Oxigraph stores.
     pub fn new(id: impl Into<String>) -> Result<Self, BlankNodeIdParseError> {
         let id = id.into();
         validate_blank_node_identifier(&id)?;
@@ -111,7 +107,14 @@ impl Default for BlankNode {
     /// Builds a new RDF [blank node](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node) with a unique id.
     #[inline]
     fn default() -> Self {
-        Self::new_from_unique_id(random::<u128>())
+        // We ensure the ID does not start with a number to be also valid with RDF/XML
+        loop {
+            let id = random();
+            let str = IdStr::new(id);
+            if matches!(str.as_str().as_bytes().first(), Some(b'a'..=b'f')) {
+                return Self(BlankNodeContent::Anonymous { id, str });
+            }
+        }
     }
 }
 
@@ -126,10 +129,7 @@ impl Default for BlankNode {
 /// ```
 /// use oxrdf::BlankNodeRef;
 ///
-/// assert_eq!(
-///     "_:a122",
-///     BlankNodeRef::new("a122")?.to_string()
-/// );
+/// assert_eq!("_:a122", BlankNodeRef::new("a122")?.to_string());
 /// # Result::<_,oxrdf::BlankNodeIdParseError>::Ok(())
 /// ```
 #[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
@@ -173,7 +173,7 @@ impl<'a> BlankNodeRef<'a> {
 
     /// Returns the underlying ID of this blank node.
     #[inline]
-    pub fn as_str(self) -> &'a str {
+    pub const fn as_str(self) -> &'a str {
         match self.0 {
             BlankNodeRefContent::Named(id) => id,
             BlankNodeRefContent::Anonymous { str, .. } => str,
@@ -185,12 +185,15 @@ impl<'a> BlankNodeRef<'a> {
     /// ```
     /// use oxrdf::BlankNode;
     ///
-    /// assert_eq!(BlankNode::new_from_unique_id(128).as_ref().unique_id(), Some(128));
+    /// assert_eq!(
+    ///     BlankNode::new_from_unique_id(128).as_ref().unique_id(),
+    ///     Some(128)
+    /// );
     /// assert_eq!(BlankNode::new("foo")?.as_ref().unique_id(), None);
     /// # Result::<_,oxrdf::BlankNodeIdParseError>::Ok(())
     /// ```
     #[inline]
-    pub fn unique_id(&self) -> Option<u128> {
+    pub const fn unique_id(&self) -> Option<u128> {
         match self.0 {
             BlankNodeRefContent::Named(_) => None,
             BlankNodeRefContent::Anonymous { id, .. } => Some(id),
@@ -244,7 +247,7 @@ impl PartialEq<BlankNodeRef<'_>> for BlankNode {
     }
 }
 
-#[derive(PartialEq, Eq, Debug, Clone, Hash)]
+#[derive(PartialEq, Eq, Debug, Clone, Hash, Serialize, Deserialize)]
 struct IdStr([u8; 32]);
 
 impl IdStr {
@@ -264,7 +267,7 @@ impl IdStr {
 
 fn validate_blank_node_identifier(id: &str) -> Result<(), BlankNodeIdParseError> {
     let mut chars = id.chars();
-    let front = chars.next().ok_or(BlankNodeIdParseError {})?;
+    let front = chars.next().ok_or(BlankNodeIdParseError)?;
     match front {
         '0'..='9'
         | '_'
@@ -283,7 +286,7 @@ fn validate_blank_node_identifier(id: &str) -> Result<(), BlankNodeIdParseError>
         | '\u{F900}'..='\u{FDCF}'
         | '\u{FDF0}'..='\u{FFFD}'
         | '\u{10000}'..='\u{EFFFF}' => (),
-        _ => return Err(BlankNodeIdParseError {}),
+        _ => return Err(BlankNodeIdParseError),
     }
     for c in chars {
         match c {
@@ -309,13 +312,13 @@ fn validate_blank_node_identifier(id: &str) -> Result<(), BlankNodeIdParseError>
             | '\u{F900}'..='\u{FDCF}'
             | '\u{FDF0}'..='\u{FFFD}'
             | '\u{10000}'..='\u{EFFFF}' => (),
-            _ => return Err(BlankNodeIdParseError {}),
+            _ => return Err(BlankNodeIdParseError),
         }
     }
 
     // Could not end with a dot
     if id.ends_with('.') {
-        Err(BlankNodeIdParseError {})
+        Err(BlankNodeIdParseError)
     } else {
         Ok(())
     }
@@ -342,19 +345,12 @@ fn to_integer_id(id: &str) -> Option<u128> {
 }
 
 /// An error raised during [`BlankNode`] IDs validation.
-#[derive(Debug)]
-pub struct BlankNodeIdParseError {}
-
-impl fmt::Display for BlankNodeIdParseError {
-    #[inline]
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(f, "The blank node identifier is invalid")
-    }
-}
-
-impl Error for BlankNodeIdParseError {}
+#[derive(Debug, thiserror::Error)]
+#[error("The blank node identifier is invalid")]
+pub struct BlankNodeIdParseError;
 
 #[cfg(test)]
+#[allow(clippy::panic_in_result_fn)]
 mod tests {
     use super::*;
 
@@ -372,13 +368,13 @@ mod tests {
 
     #[test]
     fn new_validation() {
-        assert!(BlankNode::new("").is_err());
-        assert!(BlankNode::new("a").is_ok());
-        assert!(BlankNode::new("-").is_err());
-        assert!(BlankNode::new("a-").is_ok());
-        assert!(BlankNode::new(".").is_err());
-        assert!(BlankNode::new("a.").is_err());
-        assert!(BlankNode::new("a.a").is_ok());
+        BlankNode::new("").unwrap_err();
+        BlankNode::new("a").unwrap();
+        BlankNode::new("-").unwrap_err();
+        BlankNode::new("a-").unwrap();
+        BlankNode::new(".").unwrap_err();
+        BlankNode::new("a.").unwrap_err();
+        BlankNode::new("a.a").unwrap();
     }
 
     #[test]

lib/oxrdf/src/dataset.rs

@@ -20,26 +20,28 @@
 //! assert_eq!(vec![TripleRef::new(ex, ex, ex)], results);
 //!
 //! // Print
-//! assert_eq!(dataset.to_string(), "<http://example.com> <http://example.com> <http://example.com> <http://example.com> .\n");
+//! assert_eq!(
+//!     dataset.to_string(),
+//!     "<http://example.com> <http://example.com> <http://example.com> <http://example.com> .\n"
+//! );
 //! # Result::<_,Box<dyn std::error::Error>>::Ok(())
 //! ```
 //!
 //! See also [`Graph`] if you only care about plain triples.
 
 use crate::interning::*;
-use crate::SubjectRef;
 use crate::*;
 use std::collections::hash_map::DefaultHasher;
-use std::collections::BTreeSet;
-use std::collections::{HashMap, HashSet};
+use std::collections::{BTreeSet, HashMap, HashSet};
 use std::fmt;
 use std::hash::{Hash, Hasher};
 
 /// An in-memory [RDF dataset](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-dataset).
 ///
 /// It can accommodate a fairly large number of quads (in the few millions).
-/// Beware: it interns the string and does not do any garbage collection yet:
-/// if you insert and remove a lot of different terms, memory will grow without any reduction.
+///
+/// <div class="warning">It interns the strings and does not do any garbage collection yet:
+/// if you insert and remove a lot of different terms, memory will grow without any reduction.</div>
 ///
 /// Usage example:
 /// ```
@@ -61,7 +63,7 @@ use std::hash::{Hash, Hasher};
 /// assert_eq!(vec![TripleRef::new(ex, ex, ex)], results);
 /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
 /// ```
-#[derive(Debug, Default)]
+#[derive(Debug, Default, Clone)]
 pub struct Dataset {
     interner: Interner,
     gspo: BTreeSet<(
@@ -183,6 +185,7 @@ impl Dataset {
             .map(move |q| self.decode_spog(q))
     }
 
+    #[allow(clippy::map_identity)]
     fn interned_quads_for_subject(
         &self,
         subject: &InternedSubject,
@@ -293,6 +296,18 @@ impl Dataset {
             .map(|(o, s, p, g)| (s, p, o, g))
     }
 
+    pub fn quads_for_graph_name<'a, 'b>(
+        &'a self,
+        graph_name: impl Into<GraphNameRef<'b>>,
+    ) -> impl Iterator<Item = QuadRef<'a>> + 'a {
+        let graph_name = self
+            .encoded_graph_name(graph_name)
+            .unwrap_or_else(InternedGraphName::impossible);
+
+        self.interned_quads_for_graph_name(&graph_name)
+            .map(move |q| self.decode_spog(q))
+    }
+
     fn interned_quads_for_graph_name(
         &self,
         graph_name: &InternedGraphName,
@@ -489,11 +504,11 @@ impl Dataset {
         }
     }
 
-    /// Applies on the dataset the canonicalization process described in
-    /// [Canonical Forms for Isomorphic and Equivalent RDF Graphs: Algorithms for Leaning and Labelling Blank Nodes, Aidan Hogan, 2017](http://aidanhogan.com/docs/rdf-canonicalisation.pdf).
+    /// Canonicalizes the dataset by renaming blank nodes.
     ///
     /// Usage example ([Dataset isomorphism](https://www.w3.org/TR/rdf11-concepts/#dfn-dataset-isomorphism)):
     /// ```
+    /// use oxrdf::dataset::CanonicalizationAlgorithm;
     /// use oxrdf::*;
     ///
     /// let iri = NamedNodeRef::new("http://example.com")?;
@@ -511,29 +526,59 @@ impl Dataset {
     /// graph2.insert(QuadRef::new(&bnode2, iri, iri, &g2));
     ///
     /// assert_ne!(graph1, graph2);
-    /// graph1.canonicalize();
-    /// graph2.canonicalize();
+    /// graph1.canonicalize(CanonicalizationAlgorithm::Unstable);
+    /// graph2.canonicalize(CanonicalizationAlgorithm::Unstable);
     /// assert_eq!(graph1, graph2);
     /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
     /// ```
     ///
-    /// Warning 1: Blank node ids depends on the current shape of the graph. Adding a new quad might change the ids of a lot of blank nodes.
-    /// Hence, this canonization might not be suitable for diffs.
+    /// <div class="warning">Blank node ids depends on the current shape of the graph. Adding a new quad might change the ids of a lot of blank nodes.
+    /// Hence, this canonization might not be suitable for diffs.</div>
     ///
-    /// Warning 2: The canonicalization algorithm is not stable and canonical blank node ids might change between Oxigraph version.
-    ///
-    /// Warning 3: This implementation worst-case complexity is in *O(b!)* with *b* the number of blank nodes in the input dataset.
-    pub fn canonicalize(&mut self) {
-        let bnodes = self.blank_nodes();
-        let (hash, partition) =
-            self.hash_bnodes(bnodes.into_iter().map(|bnode| (bnode, 0)).collect());
-        let new_quads = self.distinguish(&hash, &partition);
+    /// <div class="warning">This implementation worst-case complexity is in *O(b!)* with *b* the number of blank nodes in the input dataset.</div>
+    pub fn canonicalize(&mut self, algorithm: CanonicalizationAlgorithm) {
+        let bnode_mapping = self.canonicalize_interned_blank_nodes(algorithm);
+        let new_quads = self.map_blank_nodes(&bnode_mapping);
         self.clear();
         for quad in new_quads {
             self.insert_encoded(quad);
         }
     }
 
+    /// Returns a map between the current dataset blank node and the canonicalized blank node
+    /// to create a canonical dataset.
+    ///
+    /// See also [`canonicalize`](Self::canonicalize).
+    pub fn canonicalize_blank_nodes(
+        &self,
+        algorithm: CanonicalizationAlgorithm,
+    ) -> HashMap<BlankNodeRef<'_>, BlankNode> {
+        self.canonicalize_interned_blank_nodes(algorithm)
+            .into_iter()
+            .map(|(from, to)| (from.decode_from(&self.interner), to))
+            .collect()
+    }
+
+    fn canonicalize_interned_blank_nodes(
+        &self,
+        algorithm: CanonicalizationAlgorithm,
+    ) -> HashMap<InternedBlankNode, BlankNode> {
+        match algorithm {
+            CanonicalizationAlgorithm::Unstable => {
+                let bnodes = self.blank_nodes();
+                let quads_per_blank_node = self.quads_per_blank_nodes();
+                let (hash, partition) = self.hash_bnodes(
+                    bnodes.into_iter().map(|bnode| (bnode, 0)).collect(),
+                    &quads_per_blank_node,
+                );
+                self.distinguish(hash, &partition, &quads_per_blank_node)
+                    .into_iter()
+                    .map(|(from, to)| (from, BlankNode::new_from_unique_id(to.into())))
+                    .collect()
+            }
+        }
+    }
+
     fn blank_nodes(&self) -> HashSet<InternedBlankNode> {
         let mut bnodes = HashSet::new();
         for (g, s, _, o) in &self.gspo {
@@ -572,107 +617,172 @@ impl Dataset {
         }
     }
 
+    fn quads_per_blank_nodes(&self) -> QuadsPerBlankNode {
+        let mut map: HashMap<_, Vec<_>> = HashMap::new();
+        for quad in &self.spog {
+            if let InternedSubject::BlankNode(bnode) = &quad.0 {
+                map.entry(*bnode).or_default().push(quad.clone());
+            }
+            #[cfg(feature = "rdf-star")]
+            if let InternedSubject::Triple(t) = &quad.0 {
+                Self::add_quad_with_quoted_triple_to_quad_per_blank_nodes_map(quad, t, &mut map);
+            }
+            if let InternedTerm::BlankNode(bnode) = &quad.2 {
+                map.entry(*bnode).or_default().push(quad.clone());
+            }
+            #[cfg(feature = "rdf-star")]
+            if let InternedTerm::Triple(t) = &quad.2 {
+                Self::add_quad_with_quoted_triple_to_quad_per_blank_nodes_map(quad, t, &mut map);
+            }
+            if let InternedGraphName::BlankNode(bnode) = &quad.3 {
+                map.entry(*bnode).or_default().push(quad.clone());
+            }
+        }
+        map
+    }
+
+    #[cfg(feature = "rdf-star")]
+    fn add_quad_with_quoted_triple_to_quad_per_blank_nodes_map(
+        quad: &(
+            InternedSubject,
+            InternedNamedNode,
+            InternedTerm,
+            InternedGraphName,
+        ),
+        triple: &InternedTriple,
+        map: &mut QuadsPerBlankNode,
+    ) {
+        if let InternedSubject::BlankNode(bnode) = &triple.subject {
+            map.entry(*bnode).or_default().push(quad.clone());
+        }
+        if let InternedSubject::Triple(t) = &triple.subject {
+            Self::add_quad_with_quoted_triple_to_quad_per_blank_nodes_map(quad, t, map);
+        }
+        if let InternedTerm::BlankNode(bnode) = &triple.object {
+            map.entry(*bnode).or_default().push(quad.clone());
+        }
+        if let InternedTerm::Triple(t) = &triple.object {
+            Self::add_quad_with_quoted_triple_to_quad_per_blank_nodes_map(quad, t, map);
+        }
+    }
+
     fn hash_bnodes(
         &self,
         mut hashes: HashMap<InternedBlankNode, u64>,
+        quads_per_blank_node: &QuadsPerBlankNode,
     ) -> (
         HashMap<InternedBlankNode, u64>,
         Vec<(u64, Vec<InternedBlankNode>)>,
     ) {
         let mut to_hash = Vec::new();
-        let mut partition: HashMap<u64, Vec<InternedBlankNode>> = HashMap::new();
-        let mut partition_len = 0;
-        loop {
-            //TODO: improve termination
-            let mut new_hashes = HashMap::new();
-            for (bnode, old_hash) in &hashes {
-                for (_, p, o, g) in
-                    self.interned_quads_for_subject(&InternedSubject::BlankNode(*bnode))
-                {
-                    to_hash.push((
-                        self.hash_named_node(*p),
-                        self.hash_term(o, &hashes),
-                        self.hash_graph_name(g, &hashes),
-                        0,
-                    ));
-                }
-                for (s, p, _, g) in self.interned_quads_for_object(&InternedTerm::BlankNode(*bnode))
-                {
-                    to_hash.push((
-                        self.hash_subject(s, &hashes),
-                        self.hash_named_node(*p),
-                        self.hash_graph_name(g, &hashes),
-                        1,
-                    ));
-                }
-                for (s, p, o, _) in
-                    self.interned_quads_for_graph_name(&InternedGraphName::BlankNode(*bnode))
-                {
-                    to_hash.push((
-                        self.hash_subject(s, &hashes),
-                        self.hash_named_node(*p),
-                        self.hash_term(o, &hashes),
-                        2,
-                    ));
-                }
-                to_hash.sort_unstable();
-                let hash = Self::hash_tuple((old_hash, &to_hash));
-                to_hash.clear();
-                new_hashes.insert(*bnode, hash);
+        let mut to_do = hashes
+            .keys()
+            .map(|bnode| (*bnode, true))
+            .collect::<HashMap<_, _>>();
+        let mut partition = HashMap::<_, Vec<_>>::with_capacity(hashes.len());
+        let mut old_partition_count = usize::MAX;
+        while old_partition_count != partition.len() {
+            old_partition_count = partition.len();
+            partition.clear();
+            let mut new_hashes = hashes.clone();
+            for bnode in hashes.keys() {
+                let hash = if to_do.contains_key(bnode) {
+                    for (s, p, o, g) in &quads_per_blank_node[bnode] {
+                        to_hash.push((
+                            self.hash_subject(s, *bnode, &hashes),
+                            self.hash_named_node(*p),
+                            self.hash_term(o, *bnode, &hashes),
+                            self.hash_graph_name(g, *bnode, &hashes),
+                        ));
+                    }
+                    to_hash.sort_unstable();
+                    let hash = Self::hash_tuple((&to_hash, hashes[bnode]));
+                    to_hash.clear();
+                    if hash == hashes[bnode] {
+                        to_do.insert(*bnode, false);
+                    } else {
+                        new_hashes.insert(*bnode, hash);
+                    }
+                    hash
+                } else {
+                    hashes[bnode]
+                };
                 partition.entry(hash).or_default().push(*bnode);
             }
-            if partition.len() == partition_len {
-                let mut partition: Vec<_> = partition.into_iter().collect();
-                partition.sort_by(|(h1, b1), (h2, b2)| (b1.len(), h1).cmp(&(b2.len(), h2)));
-                return (hashes, partition);
-            }
             hashes = new_hashes;
-            partition_len = partition.len();
-            partition.clear();
         }
+        let mut partition: Vec<_> = partition.into_iter().collect();
+        partition.sort_unstable_by(|(h1, b1), (h2, b2)| (b1.len(), h1).cmp(&(b2.len(), h2)));
+        (hashes, partition)
     }
 
     fn hash_named_node(&self, node: InternedNamedNode) -> u64 {
         Self::hash_tuple(node.decode_from(&self.interner))
     }
 
+    fn hash_blank_node(
+        node: InternedBlankNode,
+        current_blank_node: InternedBlankNode,
+        bnodes_hash: &HashMap<InternedBlankNode, u64>,
+    ) -> u64 {
+        if node == current_blank_node {
+            u64::MAX
+        } else {
+            bnodes_hash[&node]
+        }
+    }
+
     fn hash_subject(
         &self,
         node: &InternedSubject,
+        current_blank_node: InternedBlankNode,
         bnodes_hash: &HashMap<InternedBlankNode, u64>,
     ) -> u64 {
-        #[cfg(feature = "rdf-star")]
-        if let InternedSubject::Triple(triple) = node {
-            return self.hash_triple(triple, bnodes_hash);
-        }
-        if let InternedSubject::BlankNode(bnode) = node {
-            bnodes_hash[bnode]
-        } else {
-            Self::hash_tuple(node.decode_from(&self.interner))
+        match node {
+            InternedSubject::NamedNode(node) => Self::hash_tuple(node.decode_from(&self.interner)),
+            InternedSubject::BlankNode(bnode) => {
+                Self::hash_blank_node(*bnode, current_blank_node, bnodes_hash)
+            }
+            #[cfg(feature = "rdf-star")]
+            InternedSubject::Triple(triple) => {
+                self.hash_triple(triple, current_blank_node, bnodes_hash)
+            }
         }
     }
 
-    fn hash_term(&self, term: &InternedTerm, bnodes_hash: &HashMap<InternedBlankNode, u64>) -> u64 {
-        #[cfg(feature = "rdf-star")]
-        if let InternedTerm::Triple(triple) = term {
-            return self.hash_triple(triple, bnodes_hash);
-        }
-        if let InternedTerm::BlankNode(bnode) = term {
-            bnodes_hash[bnode]
-        } else {
-            Self::hash_tuple(term.decode_from(&self.interner))
+    fn hash_term(
+        &self,
+        term: &InternedTerm,
+        current_blank_node: InternedBlankNode,
+        bnodes_hash: &HashMap<InternedBlankNode, u64>,
+    ) -> u64 {
+        match term {
+            InternedTerm::NamedNode(node) => Self::hash_tuple(node.decode_from(&self.interner)),
+            InternedTerm::BlankNode(bnode) => {
+                Self::hash_blank_node(*bnode, current_blank_node, bnodes_hash)
+            }
+            InternedTerm::Literal(literal) => Self::hash_tuple(literal.decode_from(&self.interner)),
+            #[cfg(feature = "rdf-star")]
+            InternedTerm::Triple(triple) => {
+                self.hash_triple(triple, current_blank_node, bnodes_hash)
+            }
         }
     }
 
     fn hash_graph_name(
         &self,
         graph_name: &InternedGraphName,
+        current_blank_node: InternedBlankNode,
         bnodes_hash: &HashMap<InternedBlankNode, u64>,
     ) -> u64 {
-        if let InternedGraphName::BlankNode(bnode) = graph_name {
-            bnodes_hash[bnode]
-        } else {
-            Self::hash_tuple(graph_name.decode_from(&self.interner))
+        match graph_name {
+            InternedGraphName::NamedNode(node) => {
+                Self::hash_tuple(node.decode_from(&self.interner))
+            }
+            InternedGraphName::BlankNode(bnode) => {
+                Self::hash_blank_node(*bnode, current_blank_node, bnodes_hash)
+            }
+            InternedGraphName::DefaultGraph => 0,
         }
     }
 
@@ -680,12 +790,13 @@ impl Dataset {
     fn hash_triple(
         &self,
         triple: &InternedTriple,
+        current_blank_node: InternedBlankNode,
         bnodes_hash: &HashMap<InternedBlankNode, u64>,
     ) -> u64 {
         Self::hash_tuple((
-            self.hash_subject(&triple.subject, bnodes_hash),
+            self.hash_subject(&triple.subject, current_blank_node, bnodes_hash),
             self.hash_named_node(triple.predicate),
-            self.hash_term(&triple.object, bnodes_hash),
+            self.hash_term(&triple.object, current_blank_node, bnodes_hash),
         ))
     }
 
@@ -696,46 +807,43 @@ impl Dataset {
     }
 
     fn distinguish(
-        &mut self,
-        hash: &HashMap<InternedBlankNode, u64>,
+        &self,
+        hash: HashMap<InternedBlankNode, u64>,
         partition: &[(u64, Vec<InternedBlankNode>)],
-    ) -> Vec<(
-        InternedSubject,
-        InternedNamedNode,
-        InternedTerm,
-        InternedGraphName,
-    )> {
-        let b_prime = partition.iter().find_map(|(_, b)| (b.len() > 1).then(|| b));
+        quads_per_blank_node: &QuadsPerBlankNode,
+    ) -> HashMap<InternedBlankNode, u64> {
+        let b_prime = partition.iter().map(|(_, b)| b).find(|b| b.len() > 1);
         if let Some(b_prime) = b_prime {
             b_prime
                 .iter()
                 .map(|b| {
                     let mut hash_prime = hash.clone();
                     hash_prime.insert(*b, Self::hash_tuple((hash_prime[b], 22)));
-                    let (hash_prime_prime, partition_prime) = self.hash_bnodes(hash_prime);
-                    self.distinguish(&hash_prime_prime, &partition_prime)
+                    let (hash_prime_prime, partition_prime) =
+                        self.hash_bnodes(hash_prime, quads_per_blank_node);
+                    self.distinguish(hash_prime_prime, &partition_prime, quads_per_blank_node)
                 })
-                .fold(None, |a, b| {
-                    Some(if let Some(a) = a {
-                        if a <= b {
-                            a
-                        } else {
-                            b
-                        }
+                .reduce(|a, b| {
+                    let mut a_hashes = a.values().collect::<Vec<_>>();
+                    a_hashes.sort();
+                    let mut b_hashes = a.values().collect::<Vec<_>>();
+                    b_hashes.sort();
+                    if a_hashes <= b_hashes {
+                        a
                     } else {
                         b
-                    })
+                    }
                 })
                 .unwrap_or_default()
         } else {
-            self.label(hash)
+            hash
         }
     }
 
     #[allow(clippy::needless_collect)]
-    fn label(
+    fn map_blank_nodes(
         &mut self,
-        hashes: &HashMap<InternedBlankNode, u64>,
+        bnode_mapping: &HashMap<InternedBlankNode, BlankNode>,
     ) -> Vec<(
         InternedSubject,
         InternedNamedNode,
@@ -743,103 +851,81 @@ impl Dataset {
         InternedGraphName,
     )> {
         let old_quads: Vec<_> = self.spog.iter().cloned().collect();
-        let mut quads: Vec<_> = old_quads
+        old_quads
             .into_iter()
             .map(|(s, p, o, g)| {
                 (
-                    if let InternedSubject::BlankNode(bnode) = s {
-                        InternedSubject::BlankNode(self.map_bnode(bnode, hashes))
-                    } else {
-                        #[cfg(feature = "rdf-star")]
-                        {
-                            if let InternedSubject::Triple(triple) = s {
-                                InternedSubject::Triple(Box::new(InternedTriple::encoded_into(
-                                    self.label_triple(&triple, hashes).as_ref(),
-                                    &mut self.interner,
-                                )))
-                            } else {
-                                s
-                            }
+                    match s {
+                        InternedSubject::NamedNode(_) => s,
+                        InternedSubject::BlankNode(bnode) => {
+                            InternedSubject::BlankNode(InternedBlankNode::encoded_into(
+                                bnode_mapping[&bnode].as_ref(),
+                                &mut self.interner,
+                            ))
                         }
-                        #[cfg(not(feature = "rdf-star"))]
-                        {
-                            s
+                        #[cfg(feature = "rdf-star")]
+                        InternedSubject::Triple(triple) => {
+                            InternedSubject::Triple(Box::new(InternedTriple::encoded_into(
+                                self.map_triple_blank_nodes(&triple, bnode_mapping).as_ref(),
+                                &mut self.interner,
+                            )))
                         }
                     },
                     p,
-                    if let InternedTerm::BlankNode(bnode) = o {
-                        InternedTerm::BlankNode(self.map_bnode(bnode, hashes))
-                    } else {
-                        #[cfg(feature = "rdf-star")]
-                        {
-                            if let InternedTerm::Triple(triple) = o {
-                                InternedTerm::Triple(Box::new(InternedTriple::encoded_into(
-                                    self.label_triple(&triple, hashes).as_ref(),
-                                    &mut self.interner,
-                                )))
-                            } else {
-                                o
-                            }
+                    match o {
+                        InternedTerm::NamedNode(_) | InternedTerm::Literal(_) => o,
+                        InternedTerm::BlankNode(bnode) => {
+                            InternedTerm::BlankNode(InternedBlankNode::encoded_into(
+                                bnode_mapping[&bnode].as_ref(),
+                                &mut self.interner,
+                            ))
                         }
-                        #[cfg(not(feature = "rdf-star"))]
-                        {
-                            o
+                        #[cfg(feature = "rdf-star")]
+                        InternedTerm::Triple(triple) => {
+                            InternedTerm::Triple(Box::new(InternedTriple::encoded_into(
+                                self.map_triple_blank_nodes(&triple, bnode_mapping).as_ref(),
+                                &mut self.interner,
+                            )))
                         }
                     },
-                    if let InternedGraphName::BlankNode(bnode) = g {
-                        InternedGraphName::BlankNode(self.map_bnode(bnode, hashes))
-                    } else {
-                        g
+                    match g {
+                        InternedGraphName::NamedNode(_) | InternedGraphName::DefaultGraph => g,
+                        InternedGraphName::BlankNode(bnode) => {
+                            InternedGraphName::BlankNode(InternedBlankNode::encoded_into(
+                                bnode_mapping[&bnode].as_ref(),
+                                &mut self.interner,
+                            ))
+                        }
                     },
                 )
             })
-            .collect();
-        quads.sort();
-        quads
+            .collect()
     }
 
     #[cfg(feature = "rdf-star")]
-    fn label_triple(
+    fn map_triple_blank_nodes(
         &mut self,
         triple: &InternedTriple,
-        hashes: &HashMap<InternedBlankNode, u64>,
+        bnode_mapping: &HashMap<InternedBlankNode, BlankNode>,
     ) -> Triple {
         Triple {
             subject: if let InternedSubject::BlankNode(bnode) = &triple.subject {
-                Self::gen_bnode(*bnode, hashes).into()
+                bnode_mapping[bnode].clone().into()
             } else if let InternedSubject::Triple(t) = &triple.subject {
-                self.label_triple(t, hashes).into()
+                self.map_triple_blank_nodes(t, bnode_mapping).into()
             } else {
                 triple.subject.decode_from(&self.interner).into_owned()
             },
             predicate: triple.predicate.decode_from(&self.interner).into_owned(),
             object: if let InternedTerm::BlankNode(bnode) = &triple.object {
-                Self::gen_bnode(*bnode, hashes).into()
+                bnode_mapping[bnode].clone().into()
             } else if let InternedTerm::Triple(t) = &triple.object {
-                self.label_triple(t, hashes).into()
+                self.map_triple_blank_nodes(t, bnode_mapping).into()
             } else {
                 triple.object.decode_from(&self.interner).into_owned()
             },
         }
     }
-
-    fn map_bnode(
-        &mut self,
-        old_bnode: InternedBlankNode,
-        hashes: &HashMap<InternedBlankNode, u64>,
-    ) -> InternedBlankNode {
-        InternedBlankNode::encoded_into(
-            Self::gen_bnode(old_bnode, hashes).as_ref(),
-            &mut self.interner,
-        )
-    }
-
-    fn gen_bnode(
-        old_bnode: InternedBlankNode,
-        hashes: &HashMap<InternedBlankNode, u64>,
-    ) -> BlankNode {
-        BlankNode::new_from_unique_id(hashes[&old_bnode].into())
-    }
 }
 
 impl PartialEq for Dataset {
@@ -862,7 +948,7 @@ impl<'a> IntoIterator for &'a Dataset {
     type Item = QuadRef<'a>;
     type IntoIter = Iter<'a>;
 
-    fn into_iter(self) -> Iter<'a> {
+    fn into_iter(self) -> Self::IntoIter {
         self.iter()
     }
 }
@@ -1220,7 +1306,7 @@ impl<'a> IntoIterator for GraphView<'a> {
     type Item = TripleRef<'a>;
     type IntoIter = GraphViewIter<'a>;
 
-    fn into_iter(self) -> GraphViewIter<'a> {
+    fn into_iter(self) -> Self::IntoIter {
         self.iter()
     }
 }
@@ -1229,7 +1315,7 @@ impl<'a, 'b> IntoIterator for &'b GraphView<'a> {
     type Item = TripleRef<'a>;
     type IntoIter = GraphViewIter<'a>;
 
-    fn into_iter(self) -> GraphViewIter<'a> {
+    fn into_iter(self) -> Self::IntoIter {
         self.iter()
     }
 }
@@ -1431,7 +1517,7 @@ impl<'a> IntoIterator for &'a GraphViewMut<'a> {
     type Item = TripleRef<'a>;
     type IntoIter = GraphViewIter<'a>;
 
-    fn into_iter(self) -> GraphViewIter<'a> {
+    fn into_iter(self) -> Self::IntoIter {
         self.iter()
     }
 }
@@ -1462,7 +1548,7 @@ pub struct Iter<'a> {
 impl<'a> Iterator for Iter<'a> {
     type Item = QuadRef<'a>;
 
-    fn next(&mut self) -> Option<QuadRef<'a>> {
+    fn next(&mut self) -> Option<Self::Item> {
         self.inner
             .next()
             .map(|(s, p, o, g)| self.dataset.decode_spog((s, p, o, g)))
@@ -1486,9 +1572,70 @@ pub struct GraphViewIter<'a> {
 impl<'a> Iterator for GraphViewIter<'a> {
     type Item = TripleRef<'a>;
 
-    fn next(&mut self) -> Option<TripleRef<'a>> {
+    fn next(&mut self) -> Option<Self::Item> {
         self.inner
             .next()
             .map(|(_, s, p, o)| self.dataset.decode_spo((s, p, o)))
     }
 }
+
+type QuadsPerBlankNode = HashMap<
+    InternedBlankNode,
+    Vec<(
+        InternedSubject,
+        InternedNamedNode,
+        InternedTerm,
+        InternedGraphName,
+    )>,
+>;
+
+/// An algorithm used to canonicalize graph and datasets.
+///
+/// See [`Graph::canonicalize`] and [`Dataset::canonicalize`].
+#[derive(Default, Debug, Clone, Copy, Eq, PartialEq, Hash)]
+#[non_exhaustive]
+pub enum CanonicalizationAlgorithm {
+    /// The algorithm preferred by OxRDF.
+    ///
+    /// <div class="warning">The canonicalization algorithm is not stable and canonical blank node ids might change between Oxigraph version.</div>
+    #[default]
+    Unstable,
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_canon() {
+        let mut dataset = Dataset::new();
+        dataset.insert(QuadRef::new(
+            BlankNode::default().as_ref(),
+            NamedNodeRef::new_unchecked("http://ex"),
+            BlankNode::default().as_ref(),
+            GraphNameRef::DefaultGraph,
+        ));
+        dataset.insert(QuadRef::new(
+            BlankNode::default().as_ref(),
+            NamedNodeRef::new_unchecked("http://ex"),
+            BlankNode::default().as_ref(),
+            GraphNameRef::DefaultGraph,
+        ));
+        dataset.canonicalize(CanonicalizationAlgorithm::Unstable);
+        let mut dataset2 = Dataset::new();
+        dataset2.insert(QuadRef::new(
+            BlankNode::default().as_ref(),
+            NamedNodeRef::new_unchecked("http://ex"),
+            BlankNode::default().as_ref(),
+            GraphNameRef::DefaultGraph,
+        ));
+        dataset2.insert(QuadRef::new(
+            BlankNode::default().as_ref(),
+            NamedNodeRef::new_unchecked("http://ex"),
+            BlankNode::default().as_ref(),
+            GraphNameRef::DefaultGraph,
+        ));
+        dataset2.canonicalize(CanonicalizationAlgorithm::Unstable);
+        assert_eq!(dataset, dataset2);
+    }
+}

lib/oxrdf/src/graph.rs

@@ -16,12 +16,16 @@
 //! assert_eq!(vec![triple], results);
 //!
 //! // Print
-//! assert_eq!(graph.to_string(), "<http://example.com> <http://example.com> <http://example.com> .\n");
+//! assert_eq!(
+//!     graph.to_string(),
+//!     "<http://example.com> <http://example.com> <http://example.com> .\n"
+//! );
 //! # Result::<_,Box<dyn std::error::Error>>::Ok(())
 //! ```
 //!
 //! See also [`Dataset`] if you want to get support of multiple RDF graphs at the same time.
 
+pub use crate::dataset::CanonicalizationAlgorithm;
 use crate::dataset::*;
 use crate::*;
 use std::fmt;
@@ -29,8 +33,9 @@ use std::fmt;
 /// An in-memory [RDF graph](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-graph).
 ///
 /// It can accommodate a fairly large number of triples (in the few millions).
-/// Beware: it interns the string and does not do any garbage collection yet:
-/// if you insert and remove a lot of different terms, memory will grow without any reduction.
+///
+/// <div class="warning">It interns the string and does not do any garbage collection yet:
+/// if you insert and remove a lot of different terms, memory will grow without any reduction.</div>
 ///
 /// Usage example:
 /// ```
@@ -48,7 +53,7 @@ use std::fmt;
 /// assert_eq!(vec![triple], results);
 /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
 /// ```
-#[derive(Debug, Default)]
+#[derive(Debug, Default, Clone)]
 pub struct Graph {
     dataset: Dataset,
 }
@@ -179,11 +184,11 @@ impl Graph {
         self.dataset.clear()
     }
 
-    /// Applies on the graph the canonicalization process described in
-    /// [Canonical Forms for Isomorphic and Equivalent RDF Graphs: Algorithms for Leaning and Labelling Blank Nodes, Aidan Hogan, 2017](http://aidanhogan.com/docs/rdf-canonicalisation.pdf).
+    /// Canonicalizes the dataset by renaming blank nodes.
     ///
     /// Usage example ([Graph isomorphism](https://www.w3.org/TR/rdf11-concepts/#dfn-graph-isomorphism)):
     /// ```
+    /// use oxrdf::graph::CanonicalizationAlgorithm;
     /// use oxrdf::*;
     ///
     /// let iri = NamedNodeRef::new("http://example.com")?;
@@ -199,20 +204,18 @@ impl Graph {
     /// graph2.insert(TripleRef::new(&bnode2, iri, iri));
     ///
     /// assert_ne!(graph1, graph2);
-    /// graph1.canonicalize();
-    /// graph2.canonicalize();
+    /// graph1.canonicalize(CanonicalizationAlgorithm::Unstable);
+    /// graph2.canonicalize(CanonicalizationAlgorithm::Unstable);
     /// assert_eq!(graph1, graph2);
     /// # Result::<_,Box<dyn std::error::Error>>::Ok(())
     /// ```
     ///
-    /// Warning 1: Blank node ids depends on the current shape of the graph. Adding a new triple might change the ids of a lot of blank nodes.
-    /// Hence, this canonization might not be suitable for diffs.
-    ///
-    /// Warning 2: The canonicalization algorithm is not stable and canonical blank node Ids might change between Oxigraph version.
+    /// <div class="warning">Blank node ids depends on the current shape of the graph. Adding a new quad might change the ids of a lot of blank nodes.
+    /// Hence, this canonization might not be suitable for diffs.</div>
     ///
-    /// Warning 3: This implementation worst-case complexity is in *O(b!)* with *b* the number of blank nodes in the input graph.
-    pub fn canonicalize(&mut self) {
-        self.dataset.canonicalize()
+    /// <div class="warning">This implementation worst-case complexity is in *O(b!)* with *b* the number of blank nodes in the input dataset.</div>
+    pub fn canonicalize(&mut self, algorithm: CanonicalizationAlgorithm) {
+        self.dataset.canonicalize(algorithm)
     }
 }
 
@@ -228,7 +231,7 @@ impl<'a> IntoIterator for &'a Graph {
     type Item = TripleRef<'a>;
     type IntoIter = Iter<'a>;
 
-    fn into_iter(self) -> Iter<'a> {
+    fn into_iter(self) -> Self::IntoIter {
         self.iter()
     }
 }
@@ -275,7 +278,7 @@ pub struct Iter<'a> {
 impl<'a> Iterator for Iter<'a> {
     type Item = TripleRef<'a>;
 
-    fn next(&mut self) -> Option<TripleRef<'a>> {
+    fn next(&mut self) -> Option<Self::Item> {
         self.inner.next()
     }
 }

lib/oxrdf/src/interning.rs

@@ -4,10 +4,11 @@ use crate::*;
 use std::collections::hash_map::{Entry, HashMap, RandomState};
 use std::hash::{BuildHasher, Hasher};
 
-#[derive(Debug, Default)]
+#[derive(Debug, Default, Clone)]
 pub struct Interner {
     hasher: RandomState,
     string_for_hash: HashMap<u64, String, IdentityHasherBuilder>,
+    string_for_blank_node_id: HashMap<u128, String>,
     #[cfg(feature = "rdf-star")]
     triples: HashMap<InternedTriple, Triple>,
 }
@@ -100,7 +101,7 @@ impl InternedNamedNode {
         })
     }
 
-    pub fn decode_from(self, interner: &Interner) -> NamedNodeRef {
+    pub fn decode_from(self, interner: &Interner) -> NamedNodeRef<'_> {
         NamedNodeRef::new_unchecked(interner.resolve(self.id))
     }
 
@@ -120,29 +121,53 @@ impl InternedNamedNode {
 }
 
 #[derive(Eq, PartialEq, Ord, PartialOrd, Debug, Clone, Copy, Hash)]
-pub struct InternedBlankNode {
-    id: Key,
+pub enum InternedBlankNode {
+    Number { id: u128 },
+    Other { id: Key },
 }
 
 impl InternedBlankNode {
     pub fn encoded_into(blank_node: BlankNodeRef<'_>, interner: &mut Interner) -> Self {
-        Self {
-            id: interner.get_or_intern(blank_node.as_str()),
+        if let Some(id) = blank_node.unique_id() {
+            interner
+                .string_for_blank_node_id
+                .entry(id)
+                .or_insert_with(|| blank_node.as_str().into());
+            Self::Number { id }
+        } else {
+            Self::Other {
+                id: interner.get_or_intern(blank_node.as_str()),
+            }
         }
     }
 
     pub fn encoded_from(blank_node: BlankNodeRef<'_>, interner: &Interner) -> Option<Self> {
-        Some(Self {
-            id: interner.get(blank_node.as_str())?,
-        })
+        if let Some(id) = blank_node.unique_id() {
+            interner
+                .string_for_blank_node_id
+                .contains_key(&id)
+                .then_some(Self::Number { id })
+        } else {
+            Some(Self::Other {
+                id: interner.get(blank_node.as_str())?,
+            })
+        }
     }
 
-    pub fn decode_from(self, interner: &Interner) -> BlankNodeRef {
-        BlankNodeRef::new_unchecked(interner.resolve(self.id))
+    pub fn decode_from(self, interner: &Interner) -> BlankNodeRef<'_> {
+        BlankNodeRef::new_unchecked(match self {
+            Self::Number { id } => &interner.string_for_blank_node_id[&id],
+            Self::Other { id } => interner.resolve(id),
+        })
     }
 
     pub fn next(self) -> Self {
-        Self { id: self.id.next() }
+        match self {
+            Self::Number { id } => Self::Number {
+                id: id.saturating_add(1),
+            },
+            Self::Other { id } => Self::Other { id: id.next() },
+        }
     }
 }
 
@@ -467,7 +492,7 @@ impl InternedTriple {
         interner
             .triples
             .contains_key(&interned_triple)
-            .then(|| interned_triple)
+            .then_some(interned_triple)
     }
 
     pub fn next(&self) -> Self {
@@ -479,14 +504,14 @@ impl InternedTriple {
     }
 }
 
-#[derive(Default)]
-struct IdentityHasherBuilder {}
+#[derive(Default, Clone)]
+struct IdentityHasherBuilder;
 
 impl BuildHasher for IdentityHasherBuilder {
     type Hasher = IdentityHasher;
 
-    fn build_hasher(&self) -> IdentityHasher {
-        IdentityHasher::default()
+    fn build_hasher(&self) -> Self::Hasher {
+        Self::Hasher::default()
     }
 }
 

lib/oxrdf/src/lib.rs

@@ -23,7 +23,7 @@ pub use crate::named_node::{NamedNode, NamedNodeRef};
 pub use crate::parser::TermParseError;
 pub use crate::triple::{
     GraphName, GraphNameRef, NamedOrBlankNode, NamedOrBlankNodeRef, Quad, QuadRef, Subject,
-    SubjectRef, Term, TermRef, Triple, TripleRef,
+    SubjectRef, Term, TermRef, Triple, TripleRef, TryFromTermError,
 };
 pub use crate::variable::{Variable, VariableNameParseError, VariableRef};
 pub use oxilangtag::LanguageTagParseError;

lib/oxrdf/src/literal.rs

@@ -1,22 +1,20 @@
-use crate::named_node::NamedNode;
-use crate::vocab::rdf;
-use crate::vocab::xsd;
-use crate::NamedNodeRef;
+use crate::named_node::{NamedNode, NamedNodeRef};
+use crate::vocab::{rdf, xsd};
 use oxilangtag::{LanguageTag, LanguageTagParseError};
 #[cfg(feature = "oxsdatatypes")]
 use oxsdatatypes::*;
+use serde::{Deserialize, Serialize};
 use std::borrow::Cow;
 use std::fmt;
 use std::fmt::Write;
-use std::option::Option;
 
 /// An owned RDF [literal](https://www.w3.org/TR/rdf11-concepts/#dfn-literal).
 ///
 /// The default string formatter is returning an N-Triples, Turtle, and SPARQL compatible representation:
 /// ```
 /// # use oxilangtag::LanguageTagParseError;
-/// use oxrdf::Literal;
 /// use oxrdf::vocab::xsd;
+/// use oxrdf::Literal;
 ///
 /// assert_eq!(
 ///     "\"foo\\nbar\"",
@@ -24,20 +22,20 @@ use std::option::Option;
 /// );
 ///
 /// assert_eq!(
-///     "\"1999-01-01\"^^<http://www.w3.org/2001/XMLSchema#date>",
+///     r#""1999-01-01"^^<http://www.w3.org/2001/XMLSchema#date>"#,
 ///     Literal::new_typed_literal("1999-01-01", xsd::DATE).to_string()
 /// );
 ///
 /// assert_eq!(
-///     "\"foo\"@en",
+///     r#""foo"@en"#,
 ///     Literal::new_language_tagged_literal("foo", "en")?.to_string()
 /// );
 /// # Result::<(), LanguageTagParseError>::Ok(())
 /// ```
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub struct Literal(LiteralContent);
 
-#[derive(PartialEq, Eq, Debug, Clone, Hash)]
+#[derive(PartialEq, Eq, Debug, Clone, Hash, Serialize, Deserialize)]
 enum LiteralContent {
     String(String),
     LanguageTaggedString { value: String, language: String },
@@ -427,8 +425,8 @@ impl From<DayTimeDuration> for Literal {
 ///
 /// The default string formatter is returning an N-Triples, Turtle, and SPARQL compatible representation:
 /// ```
-/// use oxrdf::LiteralRef;
 /// use oxrdf::vocab::xsd;
+/// use oxrdf::LiteralRef;
 ///
 /// assert_eq!(
 ///     "\"foo\\nbar\"",
@@ -436,7 +434,7 @@ impl From<DayTimeDuration> for Literal {
 /// );
 ///
 /// assert_eq!(
-///     "\"1999-01-01\"^^<http://www.w3.org/2001/XMLSchema#date>",
+///     r#""1999-01-01"^^<http://www.w3.org/2001/XMLSchema#date>"#,
 ///     LiteralRef::new_typed_literal("1999-01-01", xsd::DATE).to_string()
 /// );
 /// ```
@@ -459,7 +457,7 @@ enum LiteralRefContent<'a> {
 impl<'a> LiteralRef<'a> {
     /// Builds an RDF [simple literal](https://www.w3.org/TR/rdf11-concepts/#dfn-simple-literal).
     #[inline]
-    pub fn new_simple_literal(value: &'a str) -> Self {
+    pub const fn new_simple_literal(value: &'a str) -> Self {
         LiteralRef(LiteralRefContent::String(value))
     }
 
@@ -482,13 +480,13 @@ impl<'a> LiteralRef<'a> {
     ///
     /// [`Literal::new_language_tagged_literal()`] is a safe version of this constructor and should be used for untrusted data.
     #[inline]
-    pub fn new_language_tagged_literal_unchecked(value: &'a str, language: &'a str) -> Self {
+    pub const fn new_language_tagged_literal_unchecked(value: &'a str, language: &'a str) -> Self {
         LiteralRef(LiteralRefContent::LanguageTaggedString { value, language })
     }
 
     /// The literal [lexical form](https://www.w3.org/TR/rdf11-concepts/#dfn-lexical-form)
     #[inline]
-    pub fn value(self) -> &'a str {
+    pub const fn value(self) -> &'a str {
         match self.0 {
             LiteralRefContent::String(value)
             | LiteralRefContent::LanguageTaggedString { value, .. }
@@ -501,7 +499,7 @@ impl<'a> LiteralRef<'a> {
     /// Language tags are defined by the [BCP47](https://tools.ietf.org/html/bcp47).
     /// They are normalized to lowercase by this implementation.
     #[inline]
-    pub fn language(self) -> Option<&'a str> {
+    pub const fn language(self) -> Option<&'a str> {
         match self.0 {
             LiteralRefContent::LanguageTaggedString { language, .. } => Some(language),
             _ => None,
@@ -513,7 +511,7 @@ impl<'a> LiteralRef<'a> {
     /// The datatype of [language-tagged string](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tagged-string) is always [rdf:langString](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tagged-string).
     /// The datatype of [simple literals](https://www.w3.org/TR/rdf11-concepts/#dfn-simple-literal) is [xsd:string](https://www.w3.org/TR/xmlschema11-2/#string).
     #[inline]
-    pub fn datatype(self) -> NamedNodeRef<'a> {
+    pub const fn datatype(self) -> NamedNodeRef<'a> {
         match self.0 {
             LiteralRefContent::String(_) => xsd::STRING,
             LiteralRefContent::LanguageTaggedString { .. } => rdf::LANG_STRING,
@@ -526,7 +524,7 @@ impl<'a> LiteralRef<'a> {
     /// It returns true if the literal is a [language-tagged string](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tagged-string)
     /// or has the datatype [xsd:string](https://www.w3.org/TR/xmlschema11-2/#string).
     #[inline]
-    pub fn is_plain(self) -> bool {
+    pub const fn is_plain(self) -> bool {
         matches!(
             self.0,
             LiteralRefContent::String(_) | LiteralRefContent::LanguageTaggedString { .. }
@@ -552,7 +550,7 @@ impl<'a> LiteralRef<'a> {
 
     /// Extract components from this literal
     #[inline]
-    pub fn destruct(self) -> (&'a str, Option<NamedNodeRef<'a>>, Option<&'a str>) {
+    pub const fn destruct(self) -> (&'a str, Option<NamedNodeRef<'a>>, Option<&'a str>) {
         match self.0 {
             LiteralRefContent::String(s) => (s, None, None),
             LiteralRefContent::LanguageTaggedString { value, language } => {
@@ -620,17 +618,22 @@ pub fn print_quoted_str(string: &str, f: &mut impl Write) -> fmt::Result {
     f.write_char('"')?;
     for c in string.chars() {
         match c {
+            '\u{08}' => f.write_str("\\b"),
+            '\t' => f.write_str("\\t"),
             '\n' => f.write_str("\\n"),
+            '\u{0C}' => f.write_str("\\f"),
             '\r' => f.write_str("\\r"),
             '"' => f.write_str("\\\""),
             '\\' => f.write_str("\\\\"),
-            c => f.write_char(c),
+            '\0'..='\u{1F}' | '\u{7F}' => write!(f, "\\u{:04X}", u32::from(c)),
+            _ => f.write_char(c),
         }?;
     }
     f.write_char('"')
 }
 
 #[cfg(test)]
+#[allow(clippy::panic_in_result_fn)]
 mod tests {
     use super::*;
 

lib/oxrdf/src/named_node.rs

@@ -1,4 +1,5 @@
 use oxiri::{Iri, IriParseError};
+use serde::{Deserialize, Serialize};
 use std::cmp::Ordering;
 use std::fmt;
 
@@ -14,7 +15,7 @@ use std::fmt;
 /// );
 /// # Result::<_,oxrdf::IriParseError>::Ok(())
 /// ```
-#[derive(Eq, PartialEq, Ord, PartialOrd, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Ord, PartialOrd, Debug, Clone, Hash, Serialize, Deserialize)]
 pub struct NamedNode {
     iri: String,
 }
@@ -216,3 +217,21 @@ impl PartialOrd<NamedNodeRef<'_>> for NamedNode {
         self.as_ref().partial_cmp(other)
     }
 }
+
+impl From<Iri<String>> for NamedNode {
+    #[inline]
+    fn from(iri: Iri<String>) -> Self {
+        Self {
+            iri: iri.into_inner(),
+        }
+    }
+}
+
+impl<'a> From<Iri<&'a str>> for NamedNodeRef<'a> {
+    #[inline]
+    fn from(iri: Iri<&'a str>) -> Self {
+        Self {
+            iri: iri.into_inner(),
+        }
+    }
+}

lib/oxrdf/src/parser.rs

@@ -6,8 +6,6 @@ use crate::{
 #[cfg(feature = "rdf-star")]
 use crate::{Subject, Triple};
 use std::char;
-use std::error::Error;
-use std::fmt;
 use std::str::{Chars, FromStr};
 
 /// This limit is set in order to avoid stack overflow error when parsing nested triples due to too many recursive calls.
@@ -23,12 +21,15 @@ impl FromStr for NamedNode {
     /// use oxrdf::NamedNode;
     /// use std::str::FromStr;
     ///
-    /// assert_eq!(NamedNode::from_str("<http://example.com>").unwrap(), NamedNode::new("http://example.com").unwrap())
+    /// assert_eq!(
+    ///     NamedNode::from_str("<http://example.com>").unwrap(),
+    ///     NamedNode::new("http://example.com").unwrap()
+    /// )
     /// ```
-    fn from_str(s: &str) -> Result<Self, TermParseError> {
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
         let (term, left) = read_named_node(s)?;
         if !left.is_empty() {
-            return Err(TermParseError::msg(
+            return Err(Self::Err::msg(
                 "Named node serialization should end with a >",
             ));
         }
@@ -45,12 +46,15 @@ impl FromStr for BlankNode {
     /// use oxrdf::BlankNode;
     /// use std::str::FromStr;
     ///
-    /// assert_eq!(BlankNode::from_str("_:ex").unwrap(), BlankNode::new("ex").unwrap())
+    /// assert_eq!(
+    ///     BlankNode::from_str("_:ex").unwrap(),
+    ///     BlankNode::new("ex").unwrap()
+    /// )
     /// ```
-    fn from_str(s: &str) -> Result<Self, TermParseError> {
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
         let (term, left) = read_blank_node(s)?;
         if !left.is_empty() {
-            return Err(TermParseError::msg(
+            return Err(Self::Err::msg(
                 "Blank node serialization should not contain whitespaces",
             ));
         }
@@ -64,21 +68,46 @@ impl FromStr for Literal {
     /// Parses a literal from its NTriples or Turtle serialization
     ///
     /// ```
-    /// use oxrdf::{Literal, NamedNode, vocab::xsd};
+    /// use oxrdf::vocab::xsd;
+    /// use oxrdf::{Literal, NamedNode};
     /// use std::str::FromStr;
     ///
-    /// assert_eq!(Literal::from_str("\"ex\\n\"").unwrap(), Literal::new_simple_literal("ex\n"));
-    /// assert_eq!(Literal::from_str("\"ex\"@en").unwrap(), Literal::new_language_tagged_literal("ex", "en").unwrap());
-    /// assert_eq!(Literal::from_str("\"2020\"^^<http://www.w3.org/2001/XMLSchema#gYear>").unwrap(), Literal::new_typed_literal("2020", NamedNode::new("http://www.w3.org/2001/XMLSchema#gYear").unwrap()));
-    /// assert_eq!(Literal::from_str("true").unwrap(), Literal::new_typed_literal("true", xsd::BOOLEAN));
-    /// assert_eq!(Literal::from_str("+122").unwrap(), Literal::new_typed_literal("+122", xsd::INTEGER));
-    /// assert_eq!(Literal::from_str("-122.23").unwrap(), Literal::new_typed_literal("-122.23", xsd::DECIMAL));
-    /// assert_eq!(Literal::from_str("-122e+1").unwrap(), Literal::new_typed_literal("-122e+1", xsd::DOUBLE));
+    /// assert_eq!(
+    ///     Literal::from_str("\"ex\\n\"").unwrap(),
+    ///     Literal::new_simple_literal("ex\n")
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("\"ex\"@en").unwrap(),
+    ///     Literal::new_language_tagged_literal("ex", "en").unwrap()
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("\"2020\"^^<http://www.w3.org/2001/XMLSchema#gYear>").unwrap(),
+    ///     Literal::new_typed_literal(
+    ///         "2020",
+    ///         NamedNode::new("http://www.w3.org/2001/XMLSchema#gYear").unwrap()
+    ///     )
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("true").unwrap(),
+    ///     Literal::new_typed_literal("true", xsd::BOOLEAN)
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("+122").unwrap(),
+    ///     Literal::new_typed_literal("+122", xsd::INTEGER)
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("-122.23").unwrap(),
+    ///     Literal::new_typed_literal("-122.23", xsd::DECIMAL)
+    /// );
+    /// assert_eq!(
+    ///     Literal::from_str("-122e+1").unwrap(),
+    ///     Literal::new_typed_literal("-122e+1", xsd::DOUBLE)
+    /// );
     /// ```
-    fn from_str(s: &str) -> Result<Self, TermParseError> {
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
         let (term, left) = read_literal(s)?;
         if !left.is_empty() {
-            return Err(TermParseError::msg("Invalid literal serialization"));
+            return Err(Self::Err::msg("Invalid literal serialization"));
         }
         Ok(term)
     }
@@ -93,17 +122,15 @@ impl FromStr for Term {
     /// use oxrdf::*;
     /// use std::str::FromStr;
     ///
-    /// assert_eq!(Term::from_str("\"ex\"").unwrap(), Literal::new_simple_literal("ex").into());
-    /// assert_eq!(Term::from_str("<< _:s <http://example.com/p> \"o\" >>").unwrap(), Triple::new(
-    ///     BlankNode::new("s").unwrap(),
-    ///     NamedNode::new("http://example.com/p").unwrap(),
-    ///     Literal::new_simple_literal("o")
-    /// ).into());
+    /// assert_eq!(
+    ///     Term::from_str("\"ex\"").unwrap(),
+    ///     Literal::new_simple_literal("ex").into()
+    /// );
     /// ```
-    fn from_str(s: &str) -> Result<Self, TermParseError> {
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
         let (term, left) = read_term(s, 0)?;
         if !left.is_empty() {
-            return Err(TermParseError::msg("Invalid term serialization"));
+            return Err(Self::Err::msg("Invalid term serialization"));
         }
         Ok(term)
     }
@@ -118,19 +145,22 @@ impl FromStr for Variable {
     /// use oxrdf::Variable;
     /// use std::str::FromStr;
     ///
-    /// assert_eq!(Variable::from_str("$foo").unwrap(), Variable::new("foo").unwrap())
+    /// assert_eq!(
+    ///     Variable::from_str("$foo").unwrap(),
+    ///     Variable::new("foo").unwrap()
+    /// )
     /// ```
-    fn from_str(s: &str) -> Result<Self, TermParseError> {
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
         if !s.starts_with('?') && !s.starts_with('$') {
-            return Err(TermParseError::msg(
+            return Err(Self::Err::msg(
                 "Variable serialization should start with ? or $",
             ));
         }
-        Self::new(&s[1..]).map_err(|error| TermParseError {
-            kind: TermParseErrorKind::Variable {
+        Self::new(&s[1..]).map_err(|error| {
+            TermParseError(TermParseErrorKind::Variable {
                 value: s.to_owned(),
                 error,
-            },
+            })
         })
     }
 }
@@ -143,11 +173,11 @@ fn read_named_node(s: &str) -> Result<(NamedNode, &str), TermParseError> {
             .ok_or_else(|| TermParseError::msg("Named node serialization should end with a >"))?;
         let (value, remain) = remain.split_at(end);
         let remain = &remain[1..];
-        let term = NamedNode::new(value).map_err(|error| TermParseError {
-            kind: TermParseErrorKind::Iri {
+        let term = NamedNode::new(value).map_err(|error| {
+            TermParseError(TermParseErrorKind::Iri {
                 value: value.to_owned(),
                 error,
-            },
+            })
         })?;
         Ok((term, remain))
     } else {
@@ -167,11 +197,11 @@ fn read_blank_node(s: &str) -> Result<(BlankNode, &str), TermParseError> {
             })
             .unwrap_or(remain.len());
         let (value, remain) = remain.split_at(end);
-        let term = BlankNode::new(value).map_err(|error| TermParseError {
-            kind: TermParseErrorKind::BlankNode {
+        let term = BlankNode::new(value).map_err(|error| {
+            TermParseError(TermParseErrorKind::BlankNode {
                 value: value.to_owned(),
                 error,
-            },
+            })
         })?;
         Ok((term, remain))
     } else {
@@ -197,11 +227,11 @@ fn read_literal(s: &str) -> Result<(Literal, &str), TermParseError> {
                         let (language, remain) = remain.split_at(end);
                         Ok((
                             Literal::new_language_tagged_literal(value, language).map_err(
-                                |error| TermParseError {
-                                    kind: TermParseErrorKind::LanguageTag {
+                                |error| {
+                                    TermParseError(TermParseErrorKind::LanguageTag {
                                         value: language.to_owned(),
                                         error,
-                                    },
+                                    })
                                 },
                             )?,
                             remain,
@@ -217,10 +247,10 @@ fn read_literal(s: &str) -> Result<(Literal, &str), TermParseError> {
                     if let Some(c) = chars.next() {
                         value.push(match c {
                             't' => '\t',
-                            'b' => '\u{8}',
+                            'b' => '\u{08}',
                             'n' => '\n',
                             'r' => '\r',
-                            'f' => '\u{C}',
+                            'f' => '\u{0C}',
                             '"' => '"',
                             '\'' => '\'',
                             '\\' => '\\',
@@ -232,7 +262,7 @@ fn read_literal(s: &str) -> Result<(Literal, &str), TermParseError> {
                         return Err(TermParseError::msg("Unexpected literal end"));
                     }
                 }
-                c => value.push(c),
+                _ => value.push(c),
             }
         }
         Err(TermParseError::msg("Unexpected literal end"))
@@ -326,7 +356,7 @@ fn read_term(s: &str, number_of_recursive_calls: usize) -> Result<(Term, &str),
                             Term::Literal(_) => {
                                 return Err(TermParseError::msg(
                                     "Literals are not allowed in subject position",
-                                ))
+                                ));
                             }
                             Term::Triple(s) => Subject::Triple(s),
                         },
@@ -370,7 +400,7 @@ fn read_hexa_char(input: &mut Chars<'_>, len: usize) -> Result<char, TermParseEr
                     _ => {
                         return Err(TermParseError::msg(
                             "Unexpected character in a unicode escape",
-                        ))
+                        ));
                     }
                 }
         } else {
@@ -381,61 +411,59 @@ fn read_hexa_char(input: &mut Chars<'_>, len: usize) -> Result<char, TermParseEr
 }
 
 /// An error raised during term serialization parsing using the [`FromStr`] trait.
-#[derive(Debug)]
-pub struct TermParseError {
-    kind: TermParseErrorKind,
-}
+#[derive(Debug, thiserror::Error)]
+#[error(transparent)]
+pub struct TermParseError(#[from] TermParseErrorKind);
 
-#[derive(Debug)]
+/// An internal error raised during term serialization parsing using the [`FromStr`] trait.
+#[derive(Debug, thiserror::Error)]
 enum TermParseErrorKind {
-    Iri {
-        error: IriParseError,
-        value: String,
-    },
+    #[error("Error while parsing the named node '{value}': {error}")]
+    Iri { error: IriParseError, value: String },
+    #[error("Error while parsing the blank node '{value}': {error}")]
     BlankNode {
         error: BlankNodeIdParseError,
         value: String,
     },
+    #[error("Error while parsing the language tag '{value}': {error}")]
     LanguageTag {
         error: LanguageTagParseError,
         value: String,
     },
+    #[error("Error while parsing the variable '{value}': {error}")]
     Variable {
         error: VariableNameParseError,
         value: String,
     },
-    Msg {
-        msg: &'static str,
-    },
+    #[error("{0}")]
+    Msg(&'static str),
 }
 
-impl fmt::Display for TermParseError {
-    #[inline]
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        match &self.kind {
-            TermParseErrorKind::Iri { error, value } => {
-                write!(f, "Error while parsing the named node '{value}': {error}")
-            }
-            TermParseErrorKind::BlankNode { error, value } => {
-                write!(f, "Error while parsing the blank node '{value}': {error}")
-            }
-            TermParseErrorKind::LanguageTag { error, value } => {
-                write!(f, "Error while parsing the language tag '{value}': {error}")
-            }
-            TermParseErrorKind::Variable { error, value } => {
-                write!(f, "Error while parsing the variable '{value}': {error}")
-            }
-            TermParseErrorKind::Msg { msg } => f.write_str(msg),
-        }
+impl TermParseError {
+    pub(crate) fn msg(msg: &'static str) -> Self {
+        Self(TermParseErrorKind::Msg(msg))
     }
 }
 
-impl Error for TermParseError {}
+#[cfg(test)]
+#[cfg(feature = "rdf-star")]
+mod tests {
+    use super::*;
 
-impl TermParseError {
-    pub(crate) fn msg(msg: &'static str) -> Self {
-        Self {
-            kind: TermParseErrorKind::Msg { msg },
-        }
+    #[test]
+    fn triple_term_parsing() {
+        assert_eq!(
+            Term::from_str("\"ex\"").unwrap(),
+            Literal::new_simple_literal("ex").into()
+        );
+        assert_eq!(
+            Term::from_str("<< _:s <http://example.com/p> \"o\" >>").unwrap(),
+            Triple::new(
+                BlankNode::new("s").unwrap(),
+                NamedNode::new("http://example.com/p").unwrap(),
+                Literal::new_simple_literal("o"),
+            )
+            .into()
+        );
     }
 }

lib/oxrdf/src/triple.rs

@@ -2,10 +2,11 @@ use crate::blank_node::BlankNode;
 use crate::literal::Literal;
 use crate::named_node::NamedNode;
 use crate::{BlankNodeRef, LiteralRef, NamedNodeRef};
+use serde::{Deserialize, Serialize};
 use std::fmt;
 
 /// The owned union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri) and [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node).
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub enum NamedOrBlankNode {
     NamedNode(NamedNode),
     BlankNode(BlankNode),
@@ -152,7 +153,7 @@ impl<'a> From<NamedOrBlankNodeRef<'a>> for NamedOrBlankNode {
 }
 
 /// The owned union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri), [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node)  and [triples](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-triple) (if the `rdf-star` feature is enabled).
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub enum Subject {
     NamedNode(NamedNode),
     BlankNode(BlankNode),
@@ -382,7 +383,7 @@ impl<'a> From<&'a NamedOrBlankNode> for SubjectRef<'a> {
 
 /// An owned RDF [term](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-term)
 /// It is the union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri), [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node), [literals](https://www.w3.org/TR/rdf11-concepts/#dfn-literal) and [triples](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-triple) (if the `rdf-star` feature is enabled).
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub enum Term {
     NamedNode(NamedNode),
     BlankNode(BlankNode),
@@ -534,6 +535,72 @@ impl From<SubjectRef<'_>> for Term {
     }
 }
 
+impl TryFrom<Term> for NamedNode {
+    type Error = TryFromTermError;
+
+    #[inline]
+    fn try_from(term: Term) -> Result<Self, Self::Error> {
+        if let Term::NamedNode(node) = term {
+            Ok(node)
+        } else {
+            Err(TryFromTermError {
+                term,
+                target: "NamedNode",
+            })
+        }
+    }
+}
+
+impl TryFrom<Term> for BlankNode {
+    type Error = TryFromTermError;
+
+    #[inline]
+    fn try_from(term: Term) -> Result<Self, Self::Error> {
+        if let Term::BlankNode(node) = term {
+            Ok(node)
+        } else {
+            Err(TryFromTermError {
+                term,
+                target: "BlankNode",
+            })
+        }
+    }
+}
+
+impl TryFrom<Term> for Literal {
+    type Error = TryFromTermError;
+
+    #[inline]
+    fn try_from(term: Term) -> Result<Self, Self::Error> {
+        if let Term::Literal(node) = term {
+            Ok(node)
+        } else {
+            Err(TryFromTermError {
+                term,
+                target: "Literal",
+            })
+        }
+    }
+}
+
+impl TryFrom<Term> for Subject {
+    type Error = TryFromTermError;
+
+    #[inline]
+    fn try_from(term: Term) -> Result<Self, Self::Error> {
+        match term {
+            Term::NamedNode(term) => Ok(Self::NamedNode(term)),
+            Term::BlankNode(term) => Ok(Self::BlankNode(term)),
+            #[cfg(feature = "rdf-star")]
+            Term::Triple(term) => Ok(Self::Triple(term)),
+            Term::Literal(_) => Err(TryFromTermError {
+                term,
+                target: "Subject",
+            }),
+        }
+    }
+}
+
 /// A borrowed RDF [term](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-term)
 /// It is the union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri), [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node), [literals](https://www.w3.org/TR/rdf11-concepts/#dfn-literal) and [triples](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-triple) (if the `rdf-star` feature is enabled).
 #[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
@@ -698,7 +765,7 @@ impl<'a> From<TermRef<'a>> for Term {
 ///
 /// The default string formatter is returning an N-Triples, Turtle, and SPARQL compatible representation:
 /// ```
-/// use oxrdf::{Triple, NamedNode};
+/// use oxrdf::{NamedNode, Triple};
 ///
 /// assert_eq!(
 ///     "<http://example.com/s> <http://example.com/p> <http://example.com/o>",
@@ -706,11 +773,12 @@ impl<'a> From<TermRef<'a>> for Term {
 ///         subject: NamedNode::new("http://example.com/s")?.into(),
 ///         predicate: NamedNode::new("http://example.com/p")?,
 ///         object: NamedNode::new("http://example.com/o")?.into(),
-///     }.to_string()
+///     }
+///     .to_string()
 /// );
 /// # Result::<_,oxrdf::IriParseError>::Ok(())
 /// ```
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub struct Triple {
     /// The [subject](https://www.w3.org/TR/rdf11-concepts/#dfn-subject) of this triple.
     pub subject: Subject,
@@ -737,6 +805,22 @@ impl Triple {
         }
     }
 
+    /// Builds an RDF [triple](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-triple) from [`Term`]s.
+    ///
+    /// Returns a [`TryFromTermError`] error if the generated triple would be ill-formed.
+    #[inline]
+    pub fn from_terms(
+        subject: impl Into<Term>,
+        predicate: impl Into<Term>,
+        object: impl Into<Term>,
+    ) -> Result<Self, TryFromTermError> {
+        Ok(Self {
+            subject: subject.into().try_into()?,
+            predicate: predicate.into().try_into()?,
+            object: object.into(),
+        })
+    }
+
     /// Encodes that this triple is in an [RDF dataset](https://www.w3.org/TR/rdf11-concepts/#dfn-rdf-dataset).
     #[inline]
     pub fn in_graph(self, graph_name: impl Into<GraphName>) -> Quad {
@@ -769,7 +853,7 @@ impl fmt::Display for Triple {
 ///
 /// The default string formatter is returning an N-Triples, Turtle, and SPARQL compatible representation:
 /// ```
-/// use oxrdf::{TripleRef, NamedNodeRef};
+/// use oxrdf::{NamedNodeRef, TripleRef};
 ///
 /// assert_eq!(
 ///     "<http://example.com/s> <http://example.com/p> <http://example.com/o>",
@@ -777,7 +861,8 @@ impl fmt::Display for Triple {
 ///         subject: NamedNodeRef::new("http://example.com/s")?.into(),
 ///         predicate: NamedNodeRef::new("http://example.com/p")?,
 ///         object: NamedNodeRef::new("http://example.com/o")?.into(),
-///     }.to_string()
+///     }
+///     .to_string()
 /// );
 /// # Result::<_,oxrdf::IriParseError>::Ok(())
 /// ```
@@ -853,10 +938,11 @@ impl<'a> From<TripleRef<'a>> for Triple {
 
 /// A possible owned graph name.
 /// It is the union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri), [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node), and the [default graph name](https://www.w3.org/TR/rdf11-concepts/#dfn-default-graph).
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Default, Serialize, Deserialize)]
 pub enum GraphName {
     NamedNode(NamedNode),
     BlankNode(BlankNode),
+    #[default]
     DefaultGraph,
 }
 
@@ -940,10 +1026,11 @@ impl From<NamedOrBlankNodeRef<'_>> for GraphName {
 
 /// A possible borrowed graph name.
 /// It is the union of [IRIs](https://www.w3.org/TR/rdf11-concepts/#dfn-iri), [blank nodes](https://www.w3.org/TR/rdf11-concepts/#dfn-blank-node), and the [default graph name](https://www.w3.org/TR/rdf11-concepts/#dfn-default-graph).
-#[derive(Eq, PartialEq, Debug, Clone, Copy, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Copy, Hash, Default)]
 pub enum GraphNameRef<'a> {
     NamedNode(NamedNodeRef<'a>),
     BlankNode(BlankNodeRef<'a>),
+    #[default]
     DefaultGraph,
 }
 
@@ -979,7 +1066,7 @@ impl fmt::Display for GraphNameRef<'_> {
         match self {
             Self::NamedNode(node) => node.fmt(f),
             Self::BlankNode(node) => node.fmt(f),
-            Self::DefaultGraph => write!(f, "DEFAULT"),
+            Self::DefaultGraph => f.write_str("DEFAULT"),
         }
     }
 }
@@ -1060,7 +1147,7 @@ impl<'a> From<GraphNameRef<'a>> for GraphName {
 /// );
 /// # Result::<_,oxrdf::IriParseError>::Ok(())
 /// ```
-#[derive(Eq, PartialEq, Debug, Clone, Hash)]
+#[derive(Eq, PartialEq, Debug, Clone, Hash, Serialize, Deserialize)]
 pub struct Quad {
     /// The [subject](https://www.w3.org/TR/rdf11-concepts/#dfn-subject) of this triple.
     pub subject: Subject,
@@ -1220,3 +1307,62 @@ impl<'a> From<QuadRef<'a>> for Quad {
         quad.into_owned()
     }
 }
+
+/// An error return by some [`TryFrom<Term>`](TryFrom)  implementations.
+#[derive(Debug, Clone, thiserror::Error)]
+#[error("{term} can not be converted to a {target}")]
+pub struct TryFromTermError {
+    pub(crate) term: Term,
+    pub(crate) target: &'static str,
+}
+
+impl TryFromTermError {
+    /// The term that can't be converted
+    #[inline]
+    pub fn into_term(self) -> Term {
+        self.term
+    }
+}
+
+#[cfg(test)]
+#[allow(clippy::panic_in_result_fn)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn triple_from_terms() -> Result<(), TryFromTermError> {
+        assert_eq!(
+            Triple::from_terms(
+                NamedNode::new_unchecked("http://example.com/s"),
+                NamedNode::new_unchecked("http://example.com/p"),
+                NamedNode::new_unchecked("http://example.com/o"),
+            )?,
+            Triple::new(
+                NamedNode::new_unchecked("http://example.com/s"),
+                NamedNode::new_unchecked("http://example.com/p"),
+                NamedNode::new_unchecked("http://example.com/o"),
+            )
+        );
+        assert_eq!(
+            Triple::from_terms(
+                Literal::new_simple_literal("foo"),
+                NamedNode::new_unchecked("http://example.com/p"),
+                NamedNode::new_unchecked("http://example.com/o"),
+            )
+            .unwrap_err()
+            .into_term(),
+            Term::from(Literal::new_simple_literal("foo"))
+        );
+        assert_eq!(
+            Triple::from_terms(
+                NamedNode::new_unchecked("http://example.com/s"),
+                Literal::new_simple_literal("foo"),
+                NamedNode::new_unchecked("http://example.com/o"),
+            )
+            .unwrap_err()
+            .into_term(),
+            Term::from(Literal::new_simple_literal("foo"))
+        );
+        Ok(())
+    }
+}

lib/oxrdf/src/variable.rs

@@ -1,5 +1,4 @@
 use std::cmp::Ordering;
-use std::error::Error;
 use std::fmt;
 
 /// A [SPARQL query](https://www.w3.org/TR/sparql11-query/) owned variable.
@@ -8,10 +7,7 @@ use std::fmt;
 /// ```
 /// use oxrdf::{Variable, VariableNameParseError};
 ///
-/// assert_eq!(
-///     "?foo",
-///     Variable::new("foo")?.to_string()
-/// );
+/// assert_eq!("?foo", Variable::new("foo")?.to_string());
 /// # Result::<_,VariableNameParseError>::Ok(())
 /// ```
 #[derive(Eq, PartialEq, Ord, PartialOrd, Debug, Clone, Hash)]
@@ -67,12 +63,9 @@ impl fmt::Display for Variable {
 ///
 /// The default string formatter is returning a SPARQL compatible representation:
 /// ```
-/// use oxrdf::{VariableRef, VariableNameParseError};
+/// use oxrdf::{VariableNameParseError, VariableRef};
 ///
-/// assert_eq!(
-///     "?foo",
-///     VariableRef::new("foo")?.to_string()
-/// );
+/// assert_eq!("?foo", VariableRef::new("foo")?.to_string());
 /// # Result::<_,VariableNameParseError>::Ok(())
 /// ```
 #[derive(Eq, PartialEq, Ord, PartialOrd, Debug, Clone, Copy, Hash)]
@@ -96,12 +89,12 @@ impl<'a> VariableRef<'a> {
     ///
     /// [`Variable::new()`] is a safe version of this constructor and should be used for untrusted data.
     #[inline]
-    pub fn new_unchecked(name: &'a str) -> Self {
+    pub const fn new_unchecked(name: &'a str) -> Self {
         Self { name }
     }
 
     #[inline]
-    pub fn as_str(&self) -> &str {
+    pub const fn as_str(self) -> &'a str {
         self.name
     }
 
@@ -169,7 +162,7 @@ impl PartialOrd<VariableRef<'_>> for Variable {
 
 fn validate_variable_identifier(id: &str) -> Result<(), VariableNameParseError> {
     let mut chars = id.chars();
-    let front = chars.next().ok_or(VariableNameParseError {})?;
+    let front = chars.next().ok_or(VariableNameParseError)?;
     match front {
         '0'..='9'
         | '_'
@@ -188,13 +181,13 @@ fn validate_variable_identifier(id: &str) -> Result<(), VariableNameParseError>
         | '\u{F900}'..='\u{FDCF}'
         | '\u{FDF0}'..='\u{FFFD}'
         | '\u{10000}'..='\u{EFFFF}' => (),
-        _ => return Err(VariableNameParseError {}),
+        _ => return Err(VariableNameParseError),
     }
     for c in chars {
         match c {
             '0'..='9'
             | '\u{00B7}'
-            | '\u{00300}'..='\u{036F}'
+            | '\u{0300}'..='\u{036F}'
             | '\u{203F}'..='\u{2040}'
             | '_'
             | 'A'..='Z'
@@ -211,21 +204,13 @@ fn validate_variable_identifier(id: &str) -> Result<(), VariableNameParseError>
             | '\u{F900}'..='\u{FDCF}'
             | '\u{FDF0}'..='\u{FFFD}'
             | '\u{10000}'..='\u{EFFFF}' => (),
-            _ => return Err(VariableNameParseError {}),
+            _ => return Err(VariableNameParseError),
         }
     }
     Ok(())
 }
 
 /// An error raised during [`Variable`] name validation.
-#[derive(Debug)]
-pub struct VariableNameParseError {}
-
-impl fmt::Display for VariableNameParseError {
-    #[inline]
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(f, "The variable name is invalid")
-    }
-}
-
-impl Error for VariableNameParseError {}
+#[derive(Debug, thiserror::Error)]
+#[error("The variable name is invalid")]
+pub struct VariableNameParseError;

lib/oxrdf/src/vocab.rs

@@ -231,3 +231,12 @@ pub mod xsd {
     pub const YEAR_MONTH_DURATION: NamedNodeRef<'_> =
         NamedNodeRef::new_unchecked("http://www.w3.org/2001/XMLSchema#yearMonthDuration");
 }
+
+pub mod geosparql {
+    //! [GeoSpatial](https://opengeospatial.github.io/ogc-geosparql/) vocabulary.
+    use crate::named_node::NamedNodeRef;
+
+    /// Geospatial datatype like `"Point({longitude} {latitude})"^^geo:wktLiteral`
+    pub const WKT_LITERAL: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.opengis.net/ont/geosparql#wktLiteral");
+}

lib/oxrdfio/Cargo.toml

@@ -0,0 +1,36 @@
+[package]
+name = "oxrdfio"
+version = "0.1.0-alpha.5"
+authors.workspace = true
+license.workspace = true
+readme = "README.md"
+keywords = ["RDF"]
+repository = "https://github.com/oxigraph/oxigraph/tree/master/lib/oxrdfxml"
+documentation = "https://docs.rs/oxrdfio"
+description = """
+Parser and serializer for various RDF formats
+"""
+edition.workspace = true
+rust-version.workspace = true
+
+[features]
+default = []
+async-tokio = ["dep:tokio", "oxrdfxml/async-tokio", "oxttl/async-tokio"]
+rdf-star = ["oxrdf/rdf-star", "oxttl/rdf-star"]
+
+[dependencies]
+oxrdf.workspace = true
+oxrdfxml.workspace = true
+oxttl.workspace = true
+thiserror.workspace = true
+tokio = { workspace = true, optional = true, features = ["io-util"] }
+
+[dev-dependencies]
+tokio = { workspace = true, features = ["rt", "macros"] }
+
+[lints]
+workspace = true
+
+[package.metadata.docs.rs]
+all-features = true
+rustdoc-args = ["--cfg", "docsrs"]

lib/oxrdfio/README.md

@@ -0,0 +1,67 @@
+OxRDF I/O
+=========
+
+[![Latest Version](https://img.shields.io/crates/v/oxrdfio.svg)](https://crates.io/crates/oxrdfio)
+[![Released API docs](https://docs.rs/oxrdfio/badge.svg)](https://docs.rs/oxrdfio)
+[![Crates.io downloads](https://img.shields.io/crates/d/oxrdfio)](https://crates.io/crates/oxrdfio)
+[![actions status](https://github.com/oxigraph/oxigraph/workflows/build/badge.svg)](https://github.com/oxigraph/oxigraph/actions)
+[![Gitter](https://badges.gitter.im/oxigraph/community.svg)](https://gitter.im/oxigraph/community)
+
+OxRDF I/O is a set of parsers and serializers for RDF.
+
+It supports:
+* [N3](https://w3c.github.io/N3/spec/) using [`oxttl`](https://crates.io/crates/oxttl)
+* [N-Quads](https://www.w3.org/TR/n-quads/) using [`oxttl`](https://crates.io/crates/oxttl)
+* [N-Triples](https://www.w3.org/TR/n-triples/) using [`oxttl`](https://crates.io/crates/oxttl)
+* [RDF/XML](https://www.w3.org/TR/rdf-syntax-grammar/) using [`oxrdfxml`](https://crates.io/crates/oxrdfxml)
+* [TriG](https://www.w3.org/TR/trig/) using [`oxttl`](https://crates.io/crates/oxttl)
+* [Turtle](https://www.w3.org/TR/turtle/) using [`oxttl`](https://crates.io/crates/oxttl)
+
+Support for [SPARQL-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html) is also available behind the `rdf-star`feature for [Turtle-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html#turtle-star), [TriG-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html#trig-star), [N-Triples-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html#n-triples-star) and [N-Quads-star](https://w3c.github.io/rdf-star/cg-spec/2021-12-17.html#n-quads-star).
+
+It is designed as a low level parser compatible with both synchronous and asynchronous I/O (behind the `async-tokio` feature).
+
+The entry points of this library are the two [`RdfParser`] and [`RdfSerializer`] structs.
+
+Usage example converting a Turtle file to a N-Triples file:
+```rust
+use oxrdfio::{RdfFormat, RdfParser, RdfSerializer};
+
+let turtle_file = b"@base <http://example.com/> .
+@prefix schema: <http://schema.org/> .
+<foo> a schema:Person ;
+    schema:name \"Foo\" .
+<bar> a schema:Person ;
+    schema:name \"Bar\" .";
+
+let ntriples_file = b"<http://example.com/foo> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Person> .
+<http://example.com/foo> <http://schema.org/name> \"Foo\" .
+<http://example.com/bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Person> .
+<http://example.com/bar> <http://schema.org/name> \"Bar\" .
+";
+
+let mut writer = RdfSerializer::from_format(RdfFormat::NTriples).serialize_to_write(Vec::new());
+for quad in RdfParser::from_format(RdfFormat::Turtle).parse_read(turtle_file.as_ref()) {
+    writer.write_quad(&quad.unwrap()).unwrap();
+}
+assert_eq!(writer.finish().unwrap(), ntriples_file);
+```
+
+Parsers for other RDF formats exists in Rust like [graph-rdfa-processor](https://github.com/nbittich/graph-rdfa-processor) for RDFa and [json-ld](https://github.com/timothee-haudebourg/json-ld) for JSON-LD.
+
+
+## License
+
+This project is licensed under either of
+
+* Apache License, Version 2.0, ([LICENSE-APACHE](../LICENSE-APACHE) or
+  `<http://www.apache.org/licenses/LICENSE-2.0>`)
+* MIT license ([LICENSE-MIT](../LICENSE-MIT) or
+  `<http://opensource.org/licenses/MIT>`)
+
+at your option.
+
+
+### Contribution
+
+Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in Oxigraph by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.