use crate::model::{PyQuad, PyTriple};
use crate::store_utils::map_io_err;
use oxigraph::io::read::{QuadReader, TripleReader};
use oxigraph::io::{
DatasetFormat, DatasetParser, DatasetSerializer, GraphFormat, GraphParser, GraphSerializer,
};
use pyo3::exceptions::ValueError;
use pyo3::prelude::*;
use pyo3::types::PyBytes;
use pyo3::wrap_pyfunction;
use pyo3::PyIterProtocol;
use std::io;
use std::io::{BufReader, Read, Write};
pub fn add_to_module(module: &PyModule) -> PyResult<()> {
module.add_wrapped(wrap_pyfunction!(parse))?;
module.add_wrapped(wrap_pyfunction!(serialize))
}
/// Parses RDF graph and dataset serialization formats
///
/// It currently supports the following formats:
///
/// * `N-Triples `_ (``application/n-triples``)
/// * `N-Quads `_ (``application/n-quads``)
/// * `Turtle `_ (``text/turtle``)
/// * `TriG `_ (``application/trig``)
/// * `RDF/XML `_ (``application/rdf+xml``)
///
/// It supports also some MIME type aliases.
/// For example ``application/turtle`` could also be used for `Turtle `_
/// and ``application/xml`` for `RDF/XML `_.
///
/// :param input: The binary I/O object to read from. For example, it could be a file opened in binary mode with ``open('my_file.ttl', 'rb')``.
/// :type input: io.RawIOBase or io.BufferedIOBase
/// :param mime_type: the MIME type of the RDF serialization
/// :type mime_type: str
/// :param base_iri: the base IRI used to resolve the relative IRIs in the file or :py:const:`None` if relative IRI resolution should not be done
/// :type base_iri: str or None, optional
/// :return: an iterator of RDF triples or quads depending on the format
/// :rtype: iter(Triple) or iter(Quad)
/// :raises ValueError: if the MIME type is not supported
/// :raises SyntaxError: if the provided data is invalid
///
/// >>> input = io.BytesIO(b'
"1" .')
/// >>> list(parse(input, "text/turtle", base_iri="http://example.com/"))
/// [ predicate= object=>>]
#[pyfunction]
#[text_signature = "(input, /, mime_type, *, base_iri = None)"]
pub fn parse(
input: &PyAny,
mime_type: &str,
base_iri: Option<&str>,
py: Python<'_>,
) -> PyResult {
let input = BufReader::new(PyFileLike::new(input.to_object(py)));
if let Some(graph_format) = GraphFormat::from_media_type(mime_type) {
let mut parser = GraphParser::from_format(graph_format);
if let Some(base_iri) = base_iri {
parser = parser
.with_base_iri(base_iri)
.map_err(|e| ValueError::py_err(e.to_string()))?;
}
Ok(PyTripleReader {
inner: parser.read_triples(input).map_err(map_io_err)?,
}
.into_py(py))
} else if let Some(dataset_format) = DatasetFormat::from_media_type(mime_type) {
let mut parser = DatasetParser::from_format(dataset_format);
if let Some(base_iri) = base_iri {
parser = parser
.with_base_iri(base_iri)
.map_err(|e| ValueError::py_err(e.to_string()))?;
}
Ok(PyQuadReader {
inner: parser.read_quads(input).map_err(map_io_err)?,
}
.into_py(py))
} else {
Err(ValueError::py_err(format!(
"Not supported MIME type: {}",
mime_type
)))
}
}
/// Serializes an RDF graph or dataset
///
/// It currently supports the following formats:
///
/// * `N-Triples `_ (``application/n-triples``)
/// * `N-Quads `_ (``application/n-quads``)
/// * `Turtle `_ (``text/turtle``)
/// * `TriG `_ (``application/trig``)
/// * `RDF/XML `_ (``application/rdf+xml``)
///
/// It supports also some MIME type aliases.
/// For example ``application/turtle`` could also be used for `Turtle `_
/// and ``application/xml`` for `RDF/XML `_.
///
/// :param input: the RDF triples and quads to serialize
/// :type input: iter(Triple) or iter(Quad)
/// :param output: The binary I/O object to write to. For example, it could be a file opened in binary mode with ``open('my_file.ttl', 'wb')``.
/// :type output: io.RawIOBase or io.BufferedIOBase
/// :param mime_type: the MIME type of the RDF serialization
/// :type mime_type: str
/// :raises ValueError: if the MIME type is not supported
/// :raises TypeError: if a triple is given during a quad format serialization or reverse
///
/// >>> output = io.BytesIO()
/// >>> serialize([Triple(NamedNode('http://example.com'), NamedNode('http://example.com/p'), Literal('1'))], output, "text/turtle")
/// >>> output.getvalue()
/// b' "1" .\n'
#[pyfunction]
#[text_signature = "(input, output, /, mime_type, *, base_iri = None)"]
pub fn serialize(input: &PyAny, output: &PyAny, mime_type: &str, py: Python<'_>) -> PyResult<()> {
let output = PyFileLike::new(output.to_object(py));
if let Some(graph_format) = GraphFormat::from_media_type(mime_type) {
let mut writer = GraphSerializer::from_format(graph_format)
.triple_writer(output)
.map_err(map_io_err)?;
for i in input.iter()? {
writer
.write(&*i?.downcast::>()?.borrow())
.map_err(map_io_err)?;
}
writer.finish().map_err(map_io_err)?;
Ok(())
} else if let Some(dataset_format) = DatasetFormat::from_media_type(mime_type) {
let mut writer = DatasetSerializer::from_format(dataset_format)
.quad_writer(output)
.map_err(map_io_err)?;
for i in input.iter()? {
writer
.write(&*i?.downcast::>()?.borrow())
.map_err(map_io_err)?;
}
writer.finish().map_err(map_io_err)?;
Ok(())
} else {
Err(ValueError::py_err(format!(
"Not supported MIME type: {}",
mime_type
)))
}
}
#[pyclass(unsendable, name= TripleReader)]
pub struct PyTripleReader {
inner: TripleReader>,
}
#[pyproto]
impl PyIterProtocol for PyTripleReader {
fn __iter__(slf: PyRefMut) -> Py {
slf.into()
}
fn __next__(mut slf: PyRefMut) -> PyResult