Adds JSON deserializer

Closes #47

lib/Cargo.toml

@@ -45,6 +45,7 @@ sophia_api = { version = "0.6.2", optional = true }
 http = "0.2"
 httparse = { version = "1", optional = true }
 native-tls = { version = "0.2", optional = true }
+json-event-parser = "0.1"
 
 [target.'cfg(target_arch = "wasm32")'.dependencies]
 js-sys = "0.3"

lib/src/sparql/json_results.rs

@@ -1,80 +1,89 @@
 //! Implementation of [SPARQL Query Results JSON Format](https://www.w3.org/TR/sparql11-results-json/)
 
-use crate::error::invalid_input_error;
+use crate::error::{invalid_data_error, invalid_input_error};
 use crate::model::*;
 use crate::sparql::error::EvaluationError;
 use crate::sparql::model::*;
-use std::io::Write;
+use json_event_parser::{JsonEvent, JsonReader, JsonWriter};
+use std::collections::BTreeMap;
+use std::io;
+use std::io::{BufRead, Write};
+use std::rc::Rc;
 
-pub fn write_json_results(
+pub fn write_json_results(results: QueryResults, sink: impl Write) -> Result<(), EvaluationError> {
-    results: QueryResults,
+    let mut writer = JsonWriter::from_writer(sink);
-    mut sink: impl Write,
-) -> Result<(), EvaluationError> {
     match results {
         QueryResults::Boolean(value) => {
-            sink.write_all(b"{\"head\":{},\"boolean\":")?;
+            writer.write_event(JsonEvent::StartObject)?;
-            sink.write_all(if value { b"true" } else { b"false" })?;
+            writer.write_event(JsonEvent::ObjectKey("head"))?;
-            sink.write_all(b"}")?;
+            writer.write_event(JsonEvent::StartObject)?;
+            writer.write_event(JsonEvent::EndObject)?;
+            writer.write_event(JsonEvent::ObjectKey("boolean"))?;
+            writer.write_event(JsonEvent::Boolean(value))?;
+            writer.write_event(JsonEvent::EndObject)?;
             Ok(())
         }
         QueryResults::Solutions(solutions) => {
-            sink.write_all(b"{\"head\":{\"vars\":[")?;
+            writer.write_event(JsonEvent::StartObject)?;
-            let mut start_vars = true;
+            writer.write_event(JsonEvent::ObjectKey("head"))?;
+            writer.write_event(JsonEvent::StartObject)?;
+            writer.write_event(JsonEvent::ObjectKey("vars"))?;
+            writer.write_event(JsonEvent::StartArray)?;
             for variable in solutions.variables() {
-                if start_vars {
+                writer.write_event(JsonEvent::String(variable.as_str()))?;
-                    start_vars = false;
-                } else {
-                    sink.write_all(b",")?;
-                }
-                write_escaped_json_string(variable.as_str(), &mut sink)?;
             }
-            sink.write_all(b"]},\"results\":{\"bindings\":[")?;
+            writer.write_event(JsonEvent::EndArray)?;
-            let mut start_bindings = true;
+            writer.write_event(JsonEvent::EndObject)?;
+            writer.write_event(JsonEvent::ObjectKey("results"))?;
+            writer.write_event(JsonEvent::StartObject)?;
+            writer.write_event(JsonEvent::ObjectKey("bindings"))?;
+            writer.write_event(JsonEvent::StartArray)?;
             for solution in solutions {
-                if start_bindings {
+                writer.write_event(JsonEvent::StartObject)?;
-                    start_bindings = false;
-                } else {
-                    sink.write_all(b",")?;
-                }
-                sink.write_all(b"{")?;
 
                 let solution = solution?;
-                let mut start_binding = true;
                 for (variable, value) in solution.iter() {
-                    if start_binding {
+                    writer.write_event(JsonEvent::ObjectKey(variable.as_str()))?;
-                        start_binding = false;
-                    } else {
-                        sink.write_all(b",")?;
-                    }
-                    write_escaped_json_string(variable.as_str(), &mut sink)?;
                     match value {
                         Term::NamedNode(uri) => {
-                            sink.write_all(b":{\"type\":\"uri\",\"value\":")?;
+                            writer.write_event(JsonEvent::StartObject)?;
-                            write_escaped_json_string(uri.as_str(), &mut sink)?;
+                            writer.write_event(JsonEvent::ObjectKey("type"))?;
-                            sink.write_all(b"}")?;
+                            writer.write_event(JsonEvent::String("uri"))?;
+                            writer.write_event(JsonEvent::ObjectKey("value"))?;
+                            writer.write_event(JsonEvent::String(uri.as_str()))?;
+                            writer.write_event(JsonEvent::EndObject)?;
                         }
                         Term::BlankNode(bnode) => {
-                            sink.write_all(b":{\"type\":\"bnode\",\"value\":")?;
+                            writer.write_event(JsonEvent::StartObject)?;
-                            write_escaped_json_string(bnode.as_str(), &mut sink)?;
+                            writer.write_event(JsonEvent::ObjectKey("type"))?;
-                            sink.write_all(b"}")?;
+                            writer.write_event(JsonEvent::String("bnode"))?;
+                            writer.write_event(JsonEvent::ObjectKey("value"))?;
+                            writer.write_event(JsonEvent::String(bnode.as_str()))?;
+                            writer.write_event(JsonEvent::EndObject)?;
                         }
                         Term::Literal(literal) => {
-                            sink.write_all(b":{\"type\":\"literal\",\"value\":")?;
+                            writer.write_event(JsonEvent::StartObject)?;
-                            write_escaped_json_string(literal.value(), &mut sink)?;
+                            writer.write_event(JsonEvent::ObjectKey("type"))?;
+                            writer.write_event(JsonEvent::String("literal"))?;
+                            writer.write_event(JsonEvent::ObjectKey("value"))?;
+                            writer.write_event(JsonEvent::String(literal.value()))?;
                             if let Some(language) = literal.language() {
-                                sink.write_all(b",\"xml:lang\":")?;
+                                writer.write_event(JsonEvent::ObjectKey("xml:lang"))?;
-                                write_escaped_json_string(language, &mut sink)?;
+                                writer.write_event(JsonEvent::String(language))?;
                             } else if !literal.is_plain() {
-                                sink.write_all(b",\"datatype\":")?;
+                                writer.write_event(JsonEvent::ObjectKey("datatype"))?;
-                                write_escaped_json_string(literal.datatype().as_str(), &mut sink)?;
+                                writer
+                                    .write_event(JsonEvent::String(literal.datatype().as_str()))?;
                             }
-                            sink.write_all(b"}")?;
+                            writer.write_event(JsonEvent::EndObject)?;
                         }
                     }
                 }
-                sink.write_all(b"}")?;
+                writer.write_event(JsonEvent::EndObject)?;
             }
-            sink.write_all(b"]}}")?;
+            writer.write_event(JsonEvent::EndArray)?;
+            writer.write_event(JsonEvent::EndObject)?;
+            writer.write_event(JsonEvent::EndObject)?;
             Ok(())
         }
         QueryResults::Graph(_) => Err(invalid_input_error(
@@ -84,37 +93,253 @@ pub fn write_json_results(
     }
 }
 
-fn write_escaped_json_string(s: &str, mut sink: impl Write) -> Result<(), EvaluationError> {
+pub fn read_json_results(source: impl BufRead + 'static) -> Result<QueryResults, io::Error> {
-    sink.write_all(b"\"")?;
+    let mut reader = JsonReader::from_reader(source);
-    for c in s.chars() {
+    let mut buffer = Vec::default();
-        match c {
+    let mut variables = None;
-            '\\' => sink.write_all(b"\\\\"),
+
-            '"' => sink.write_all(b"\\\""),
+    if reader.read_event(&mut buffer)? != JsonEvent::StartObject {
-            c => {
+        return Err(invalid_data_error(
-                if c < char::from(32) {
+            "SPARQL JSON results should be an object",
-                    match c {
+        ));
-                        '\u{08}' => sink.write_all(b"\\b"),
+    }
-                        '\u{0C}' => sink.write_all(b"\\f"),
+
-                        '\n' => sink.write_all(b"\\n"),
+    loop {
-                        '\r' => sink.write_all(b"\\r"),
+        let event = reader.read_event(&mut buffer)?;
-                        '\t' => sink.write_all(b"\\t"),
+        match event {
-                        c => {
+            JsonEvent::ObjectKey(key) => match key {
-                            let mut c = c as u8;
+                "head" => variables = Some(read_head(&mut reader, &mut buffer)?),
-                            let mut result = [b'\\', b'u', 0, 0, 0, 0];
+                "results" => {
-                            for i in (2..6).rev() {
+                    if reader.read_event(&mut buffer)? != JsonEvent::StartObject {
-                                let ch = c % 16;
+                        return Err(invalid_data_error("'results' should be an object"));
-                                result[i] = ch + if ch < 10 { b'0' } else { b'A' };
+                    }
-                                c /= 16;
+                    if reader.read_event(&mut buffer)? != JsonEvent::ObjectKey("bindings") {
-                            }
+                        return Err(invalid_data_error(
-                            sink.write_all(&result)
+                            "'results' should contain a 'bindings' key",
+                        ));
+                    }
+                    if reader.read_event(&mut buffer)? != JsonEvent::StartArray {
+                        return Err(invalid_data_error("'bindings' should be an object"));
                     }
+                    return if let Some(variables) = variables {
+                        let mut mapping = BTreeMap::default();
+                        for (i, var) in variables.iter().enumerate() {
+                            mapping.insert(var.clone(), i);
                         }
+                        Ok(QueryResults::Solutions(QuerySolutionIter::new(
+                            Rc::new(
+                                variables
+                                    .into_iter()
+                                    .map(Variable::new)
+                                    .collect::<Result<Vec<_>, _>>()
+                                    .map_err(invalid_data_error)?,
+                            ),
+                            Box::new(ResultsIterator {
+                                reader,
+                                buffer,
+                                mapping,
+                            }),
+                        )))
                     } else {
-                    write!(sink, "{}", c)
+                        Err(invalid_data_error(
+                            "SPARQL tuple query results should contain a head key",
+                        ))
+                    };
+                }
+                "boolean" => {
+                    return if let JsonEvent::Boolean(v) = reader.read_event(&mut buffer)? {
+                        Ok(QueryResults::Boolean(v))
+                    } else {
+                        Err(invalid_data_error("Unexpected boolean value"))
+                    }
+                }
+                _ => {
+                    return Err(invalid_data_error(format!(
+                        "Expecting head or result key, found {}",
+                        key
+                    )));
+                }
+            },
+            JsonEvent::EndObject => {
+                return Err(invalid_data_error(
+                    "SPARQL results should contain a bindings key or a boolean key",
+                ))
+            }
+            JsonEvent::Eof => return Err(io::Error::from(io::ErrorKind::UnexpectedEof)),
+            _ => return Err(invalid_data_error("Invalid SPARQL results serialization")),
+        }
+    }
+}
+
+fn read_head<R: BufRead>(
+    reader: &mut JsonReader<R>,
+    buffer: &mut Vec<u8>,
+) -> io::Result<Vec<String>> {
+    if reader.read_event(buffer)? != JsonEvent::StartObject {
+        return Err(invalid_data_error("head should be an object"));
+    }
+    let mut variables = None;
+    loop {
+        match reader.read_event(buffer)? {
+            JsonEvent::ObjectKey(key) => match key {
+                "vars" => variables = Some(read_string_array(reader, buffer)?),
+                "link" => {
+                    read_string_array(reader, buffer)?;
+                }
+                _ => {
+                    return Err(invalid_data_error(format!(
+                        "Unexpected key in head: '{}'",
+                        key
+                    )))
+                }
+            },
+            JsonEvent::EndObject => return Ok(variables.unwrap_or_else(Vec::new)),
+            _ => return Err(invalid_data_error("Invalid head serialization")),
+        }
+    }
+}
+
+fn read_string_array<R: BufRead>(
+    reader: &mut JsonReader<R>,
+    buffer: &mut Vec<u8>,
+) -> io::Result<Vec<String>> {
+    if reader.read_event(buffer)? != JsonEvent::StartArray {
+        return Err(invalid_data_error("Variable list should be an array"));
+    }
+    let mut elements = Vec::new();
+    loop {
+        match reader.read_event(buffer)? {
+            JsonEvent::String(s) => {
+                elements.push(s.into());
+            }
+            JsonEvent::EndArray => return Ok(elements),
+            _ => return Err(invalid_data_error("Variable names should be strings")),
+        }
+    }
+}
+
+struct ResultsIterator<R: BufRead> {
+    reader: JsonReader<R>,
+    buffer: Vec<u8>,
+    mapping: BTreeMap<String, usize>,
+}
+
+impl<R: BufRead> Iterator for ResultsIterator<R> {
+    type Item = Result<Vec<Option<Term>>, EvaluationError>;
+
+    fn next(&mut self) -> Option<Result<Vec<Option<Term>>, EvaluationError>> {
+        self.read_next().map_err(EvaluationError::from).transpose()
+    }
+}
+
+impl<R: BufRead> ResultsIterator<R> {
+    fn read_next(&mut self) -> io::Result<Option<Vec<Option<Term>>>> {
+        let mut new_bindings = vec![None; self.mapping.len()];
+        loop {
+            match self.reader.read_event(&mut self.buffer)? {
+                JsonEvent::StartObject => (),
+                JsonEvent::EndObject => return Ok(Some(new_bindings)),
+                JsonEvent::EndArray | JsonEvent::Eof => return Ok(None),
+                JsonEvent::ObjectKey(key) => {
+                    let k = *self.mapping.get(key).ok_or_else(|| {
+                        invalid_data_error(format!(
+                            "The variable {} has not been defined in the header",
+                            key
+                        ))
+                    })?;
+                    new_bindings[k] = Some(self.read_value()?)
+                }
+                _ => return Err(invalid_data_error("Invalid result serialization")),
+            }
+        }
+    }
+    fn read_value(&mut self) -> io::Result<Term> {
+        enum Type {
+            Uri,
+            BNode,
+            Literal,
+        }
+        enum State {
+            Type,
+            Value,
+            Lang,
+            Datatype,
+        }
+        let mut state = None;
+        let mut t = None;
+        let mut value = None;
+        let mut lang = None;
+        let mut datatype = None;
+        if self.reader.read_event(&mut self.buffer)? != JsonEvent::StartObject {
+            return Err(invalid_data_error(
+                "Term serializations should be an object",
+            ));
+        }
+        loop {
+            match self.reader.read_event(&mut self.buffer)? {
+                JsonEvent::ObjectKey(key) => match key {
+                    "type" => state = Some(State::Type),
+                    "value" => state = Some(State::Value),
+                    "xml:lang" => state = Some(State::Lang),
+                    "datatype" => state = Some(State::Datatype),
+                    _ => {
+                        return Err(invalid_data_error(format!(
+                            "Unexpected key in term serialization: '{}'",
+                            key
+                        )))
+                    }
+                },
+                JsonEvent::String(s) => match state {
+                    None => (), // impossible
+                    Some(State::Type) => match s {
+                        "uri" => t = Some(Type::Uri),
+                        "bnode" => t = Some(Type::BNode),
+                        "literal" => t = Some(Type::Literal),
+                        _ => {
+                            return Err(invalid_data_error(format!(
+                                "Unexpected term type: '{}'",
+                                s
+                            )))
+                        }
+                    },
+                    Some(State::Value) => value = Some(s.to_owned()),
+                    Some(State::Lang) => lang = Some(s.to_owned()),
+                    Some(State::Datatype) => datatype = Some(s.to_owned()),
+                },
+                JsonEvent::EndObject => {
+                    let value = value.ok_or_else(|| {
+                        invalid_data_error("Term serialization should have a value key")
+                    })?;
+                    return match t {
+                        None => Err(invalid_data_error(
+                            "Term serialization should have a type key",
+                        )),
+                        Some(Type::Uri) => Ok(NamedNode::new(value)
+                            .map_err(|e| invalid_data_error(format!("Invalid uri value: {}", e)))?
+                            .into()),
+                        Some(Type::BNode) => Ok(BlankNode::new(value)
+                            .map_err(|e| invalid_data_error(format!("Invalid bnode value: {}", e)))?
+                            .into()),
+                        Some(Type::Literal) => Ok(match datatype {
+                            Some(datatype) => Literal::new_typed_literal(
+                                value,
+                                NamedNode::new(datatype).map_err(|e| {
+                                    invalid_data_error(format!("Invalid datatype value: {}", e))
+                                })?,
+                            ),
+                            None => match lang {
+                                Some(lang) => Literal::new_language_tagged_literal(value, lang)
+                                    .map_err(|e| {
+                                        invalid_data_error(format!("Invalid xml:lang value: {}", e))
+                                    })?,
+                                None => Literal::new_simple_literal(value),
+                            },
+                        }
+                        .into()),
+                    };
+                }
+                _ => return Err(invalid_data_error("Invalid term serialization")),
             }
         }
-        }?;
     }
-    sink.write_all(b"\"")?;
-    Ok(())
 }

lib/src/sparql/model.rs

@@ -4,7 +4,7 @@ use crate::io::GraphSerializer;
 use crate::model::*;
 use crate::sparql::csv_results::{read_tsv_results, write_csv_results, write_tsv_results};
 use crate::sparql::error::EvaluationError;
-use crate::sparql::json_results::write_json_results;
+use crate::sparql::json_results::{read_json_results, write_json_results};
 use crate::sparql::xml_results::{read_xml_results, write_xml_results};
 use rand::random;
 use std::error::Error;
@@ -30,11 +30,9 @@ impl QueryResults {
     ) -> Result<Self, io::Error> {
         match format {
             QueryResultsFormat::Xml => read_xml_results(reader),
-            QueryResultsFormat::Json => Err(invalid_input_error(
+            QueryResultsFormat::Json => read_json_results(reader),
-                "JSON SPARQL results format parsing has not been implemented yet",
-            )), //TODO: implement
             QueryResultsFormat::Csv => Err(invalid_input_error(
-                "CSV and TSV SPARQL results format parsing is not implemented",
+                "CSV SPARQL results format parsing is not implemented",
             )),
             QueryResultsFormat::Tsv => read_tsv_results(reader),
         }
@@ -520,3 +518,73 @@ impl fmt::Display for VariableNameParseError {
 }
 
 impl Error for VariableNameParseError {}
+
+#[test]
+fn test_serialization_rountrip() -> Result<(), EvaluationError> {
+    use std::io::Cursor;
+    use std::str;
+
+    for format in &[
+        QueryResultsFormat::Xml,
+        QueryResultsFormat::Json,
+        QueryResultsFormat::Tsv,
+    ] {
+        let results = vec![
+            QueryResults::Boolean(true),
+            QueryResults::Boolean(false),
+            QueryResults::Solutions(QuerySolutionIter::new(
+                Rc::new(vec![
+                    Variable::new_unchecked("foo"),
+                    Variable::new_unchecked("bar"),
+                ]),
+                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()),
+                        ]),
+                    ]
+                    .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/src/sparql/xml_results.rs

@@ -318,8 +318,7 @@ impl<R: BufRead> ResultsIterator<R> {
         }
         let mut state = State::Start;
 
-        let mut new_bindings = Vec::default();
+        let mut new_bindings = vec![None; self.mapping.len()];
-        new_bindings.resize(self.mapping.len(), None);
 
         let mut current_var = None;
         let mut term: Option<Term> = None;
@@ -474,13 +473,12 @@ impl<R: BufRead> ResultsIterator<R> {
                     State::Result => return Ok(Some(new_bindings)),
                     State::Binding => {
                         if let Some(var) = &current_var {
-                            new_bindings[self.mapping[var]] = term.clone()
+                            new_bindings[self.mapping[var]] = term.take()
                         } else {
                             return Err(
                                 invalid_data_error("No name found for <binding> tag").into()
                             );
                         }
-                        term = None;
                         state = State::Result;
                     }
                     State::Uri | State::BNode => state = State::Binding,

testsuite/tests/sparql.rs

@@ -85,15 +85,12 @@ fn sparql11_query_w3c_evaluation_testsuite() -> Result<()> {
             "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/syntax-query/manifest#test_61a",
             "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/syntax-query/manifest#test_62a",
             "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/syntax-query/manifest#test_65",
-            // SPARQL 1.1 JSON query results deserialization is not implemented yet
-            "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/aggregates/manifest#agg-empty-group-count-1",
-            "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/aggregates/manifest#agg-empty-group-count-2",
             //BNODE() scope is currently wrong
             "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/functions/manifest#bnode01",
             //Property path with unbound graph name are not supported yet
             "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/property-path/manifest#pp35",
             //SERVICE name from a BGP
-            "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/service/manifest#service5"
+            "http://www.w3.org/2009/sparql/docs/tests/data-sparql11/service/manifest#service5",
         ],
     )
 }