pyjelly 0.2.3__py3-none-any.whl → 0.3.0__py3-none-any.whl

pyjelly/integrations/rdflib/parse.py

@@ -1,23 +1,30 @@
 from __future__ import annotations
 
 from collections.abc import Generator, Iterable
-from typing import IO, Any
+from typing import IO, Any, Callable
 from typing_extensions import Never, override
 
 import rdflib
 from rdflib.graph import DATASET_DEFAULT_GRAPH_ID, Dataset, Graph
 from rdflib.parser import InputSource
 from rdflib.parser import Parser as RDFLibParser
-from rdflib.store import Store
 
 from pyjelly import jelly
 from pyjelly.errors import JellyConformanceError
 from pyjelly.options import StreamTypes
-from pyjelly.parse.decode import Adapter, Decoder, ParserOptions
+from pyjelly.parse.decode import Adapter, Decoder, ParserOptions, ParsingMode
 from pyjelly.parse.ioutils import get_options_and_frames
 
 
 class RDFLibAdapter(Adapter):
+    """
+    RDFLib adapter class, is extended by triples and quads implementations.
+
+    Args:
+        Adapter (_type_): abstract adapter class
+
+    """
+
     @override
     def iri(self, iri: str) -> rdflib.URIRef:
         return rdflib.URIRef(iri)
@@ -41,6 +48,24 @@ class RDFLibAdapter(Adapter):
 
 
 def _adapter_missing(feature: str, *, stream_types: StreamTypes) -> Never:
+    """
+    Raise error if functionality is missing in adapter.
+
+    TODO: currently not used anywhere due to logical types being removed
+
+    Args:
+        feature (str): function which is not implemented
+        stream_types (StreamTypes): what combination of physical/logical types
+            triggered the error
+
+    Raises:
+        NotImplementedError: raises error with message with missing functionality
+            and types encountered
+
+    Returns:
+        Never: only raises errors
+
+    """
     physical_type_name = jelly.PhysicalStreamType.Name(stream_types.physical_type)
     logical_type_name = jelly.LogicalStreamType.Name(stream_types.logical_type)
     msg = (
@@ -51,242 +76,336 @@ def _adapter_missing(feature: str, *, stream_types: StreamTypes) -> Never:
 
 
 class RDFLibTriplesAdapter(RDFLibAdapter):
-    graph: Graph
+    """
+    Triples adapter RDFLib implementation.
+
+    Notes: has internal graph object which tracks
+        triples and namespaces and can get flushed between frames.
+    """
 
-    def __init__(self, options: ParserOptions, store: Store | str = "default") -> None:
-        super().__init__(options=options)
-        self.graph = Graph(store=store)
+    def __init__(
+        self,
+        options: ParserOptions,
+        graph_factory: Callable[[], Graph],
+        parsing_mode: ParsingMode = ParsingMode.FLAT,
+    ) -> None:
+        super().__init__(options=options, parsing_mode=parsing_mode)
+        self.graph = graph_factory()
+        self.graph_factory = graph_factory
+        self.parsing_mode = parsing_mode
 
     @override
     def triple(self, terms: Iterable[Any]) -> Any:
-        self.graph.add(terms)  # type: ignore[arg-type]
+        self.graph.add(tuple(terms))
 
     @override
     def namespace_declaration(self, name: str, iri: str) -> None:
         self.graph.bind(name, self.iri(iri))
 
-    def frame(self) -> Graph | None:
-        if self.options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_GRAPHS:
-            this_graph = self.graph
-            self.graph = Graph(store=self.graph.store)
-            return this_graph
-        if self.options.stream_types.logical_type in (
-            jelly.LOGICAL_STREAM_TYPE_UNSPECIFIED,
-            jelly.LOGICAL_STREAM_TYPE_FLAT_TRIPLES,
-        ):
-            return None
-        return _adapter_missing(
-            "interpreting frames",
-            stream_types=self.options.stream_types,
-        )
+    def frame(self) -> Graph:
+        """
+        Finalize one frame in triples stream.
+
+        Returns:
+           Graph: frame content as a separate Graph
+                and starts a new Graph
+
+        """
+        this_graph = self.graph
+        self.graph = self.graph_factory()
+        return this_graph
 
 
 class RDFLibQuadsBaseAdapter(RDFLibAdapter):
     def __init__(
         self,
         options: ParserOptions,
-        store: Store | str,
+        dataset_factory: Callable[[], Dataset],
+        parsing_mode: ParsingMode = ParsingMode.FLAT,
     ) -> None:
-        super().__init__(options=options)
-        self.store = store
-        self.dataset = self.new_dataset()
-
-    def new_dataset(self) -> Dataset:
-        return Dataset(store=self.store, default_union=True)
+        super().__init__(options=options, parsing_mode=parsing_mode)
+        self.dataset = dataset_factory()
+        self.dataset_factory = dataset_factory
 
     @override
-    def frame(self) -> Dataset | None:
-        if self.options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_DATASETS:
-            this_dataset = self.dataset
-            self.dataset = self.new_dataset()
-            return this_dataset
-        if self.options.stream_types.logical_type in (
-            jelly.LOGICAL_STREAM_TYPE_UNSPECIFIED,
-            jelly.LOGICAL_STREAM_TYPE_FLAT_QUADS,
-        ):
-            return None
-        return _adapter_missing(
-            "interpreting frames", stream_types=self.options.stream_types
-        )
+    def frame(self) -> Dataset:
+        current_dataset = self.dataset
+        self.dataset = self.dataset_factory()
+        return current_dataset
 
 
 class RDFLibQuadsAdapter(RDFLibQuadsBaseAdapter):
+    """
+    Extended RDFLib adapter for the QUADS physical type.
+
+    Notes:
+        Adds triples and namespaces directly to
+        dataset, so RDFLib handles the rest.
+
+    Args:
+        RDFLibQuadsBaseAdapter (_type_): base quads adapter
+            (shared with graphs physical type)
+
+    """
+
     @override
     def namespace_declaration(self, name: str, iri: str) -> None:
         self.dataset.bind(name, self.iri(iri))
 
     @override
     def quad(self, terms: Iterable[Any]) -> Any:
-        self.dataset.add(terms)  # type: ignore[arg-type]
+        self.dataset.add(tuple(terms))
 
 
 class RDFLibGraphsAdapter(RDFLibQuadsBaseAdapter):
-    _graph: Graph | None = None
+    """
+    Extension of RDFLibQuadsBaseAdapter for the GRAPHS physical type.
+
+    Notes: introduces graph start/end, checks if graph exists,
+        dataset store management.
+
+    Args:
+        RDFLibQuadsBaseAdapter (_type_): base adapter for quads management.
+
+    Raises:
+        JellyConformanceError: if no graph_start was encountered
+
+    """
+
+    _graph_id: str | None
 
     def __init__(
         self,
         options: ParserOptions,
-        store: Store | str,
+        dataset_factory: Callable[[], Dataset],
+        parsing_mode: ParsingMode = ParsingMode.FLAT,
     ) -> None:
-        super().__init__(options=options, store=store)
-        self._graph = None
+        super().__init__(
+            options=options,
+            dataset_factory=dataset_factory,
+            parsing_mode=parsing_mode,
+        )
+        self._graph_id = None
 
     @property
-    def graph(self) -> Graph:
-        if self._graph is None:
+    def graph(self) -> None:
+        if self._graph_id is None:
             msg = "new graph was not started"
             raise JellyConformanceError(msg)
-        return self._graph
 
     @override
     def graph_start(self, graph_id: str) -> None:
-        self._graph = Graph(store=self.dataset.store, identifier=graph_id)
+        self._graph_id = graph_id
 
     @override
     def namespace_declaration(self, name: str, iri: str) -> None:
-        self.graph.bind(name, self.iri(iri))
+        self.dataset.bind(name, self.iri(iri))
 
     @override
     def triple(self, terms: Iterable[Any]) -> None:
-        self.graph.add(terms)  # type: ignore[arg-type]
+        self.dataset.add((*terms, self._graph_id))
 
     @override
     def graph_end(self) -> None:
-        self.dataset.store.add_graph(self.graph)
-        self._graph = None
+        self._graph_id = None
 
-    def frame(self) -> Dataset | None:
-        if self.options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_DATASETS:
-            this_dataset = self.dataset
-            self._graph = None
-            self.dataset = self.new_dataset()
-            return this_dataset
-        return super().frame()
 
-
-def parse_flat_triples_stream(
+def parse_triples_stream(
     frames: Iterable[jelly.RdfStreamFrame],
     options: ParserOptions,
-    store: Store | str = "default",
-    identifier: str | None = None,
-) -> Dataset | Graph:
-    assert options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_TRIPLES
-    adapter = RDFLibTriplesAdapter(options, store=store)
-    if identifier is not None:
-        adapter.graph = Graph(identifier=identifier, store=store)
+    graph_factory: Callable[[], Graph],
+    parsing_mode: ParsingMode = ParsingMode.FLAT,
+) -> Generator[Graph]:
+    """
+    Parse flat triple stream.
+
+    Args:
+        frames (Iterable[jelly.RdfStreamFrame]): iterator over stream frames
+        options (ParserOptions): stream options
+        graph_factory (Callable): Lambda to construct a graph
+        parsing_mode (ParsingMode): specifies whether this is
+            a flat or grouped parsing.
+
+    Yields:
+        Generator[Graph]: RDFLib Graph(s)
+
+    """
+    adapter = RDFLibTriplesAdapter(
+        options, graph_factory=graph_factory, parsing_mode=parsing_mode
+    )
     decoder = Decoder(adapter=adapter)
     for frame in frames:
-        decoder.decode_frame(frame=frame)
-    return adapter.graph
+        g = decoder.decode_frame(frame)
+        if g is not None:
+            yield g
+
+    if parsing_mode is ParsingMode.FLAT:
+        yield adapter.graph
 
 
-def parse_flat_quads_stream(
+def parse_quads_stream(
     frames: Iterable[jelly.RdfStreamFrame],
     options: ParserOptions,
-    store: Store | str = "default",
-    identifier: str | None = None,
-) -> Dataset:
-    assert options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_QUADS
+    dataset_factory: Callable[[], Dataset],
+    parsing_mode: ParsingMode = ParsingMode.FLAT,
+) -> Generator[Dataset]:
+    """
+    Parse flat quads stream.
+
+    Args:
+        frames (Iterable[jelly.RdfStreamFrame]): iterator over stream frames
+        options (ParserOptions): stream options
+        dataset_factory (Callable): Lambda to construct a dataset
+        parsing_mode (ParsingMode): specifies whether this is
+            a flat or grouped parsing.
+
+    Yields:
+        Generator[Dataset]: RDFLib dataset(s)
+
+    """
     adapter_class: type[RDFLibQuadsBaseAdapter]
     if options.stream_types.physical_type == jelly.PHYSICAL_STREAM_TYPE_QUADS:
         adapter_class = RDFLibQuadsAdapter
-    else:  # jelly.PHYSICAL_STREAM_TYPE_GRAPHS
+    else:
         adapter_class = RDFLibGraphsAdapter
-    adapter = adapter_class(options=options, store=store)
-    adapter.dataset.default_context = Graph(identifier=identifier, store=store)
+    adapter = adapter_class(
+        options=options,
+        dataset_factory=dataset_factory,
+        parsing_mode=parsing_mode,
+    )
     decoder = Decoder(adapter=adapter)
     for frame in frames:
-        decoder.decode_frame(frame=frame)
-    return adapter.dataset
+        ds = decoder.decode_frame(frame)
+        if ds is not None:
+            yield ds
 
-
-def parse_graph_stream(
-    frames: Iterable[jelly.RdfStreamFrame],
-    options: ParserOptions,
-    store: Store | str = "default",
-) -> Generator[Graph]:
-    assert options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_GRAPHS
-    adapter = RDFLibTriplesAdapter(options, store=store)
-    decoder = Decoder(adapter=adapter)
-    for frame in frames:
-        yield decoder.decode_frame(frame=frame)
+    if parsing_mode is ParsingMode.FLAT:
+        yield adapter.dataset
 
 
-def graphs_from_jelly(
+def parse_jelly_grouped(
     inp: IO[bytes],
-    store: Store | str = "default",
-) -> Generator[Any] | Generator[Dataset] | Generator[Graph]:
-    options, frames = get_options_and_frames(inp)
-
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_TRIPLES:
-        yield parse_flat_triples_stream(frames=frames, options=options, store=store)
-        return
+    graph_factory: Callable[[], Graph],
+    dataset_factory: Callable[[], Dataset],
+) -> Generator[Any] | Generator[Graph] | Generator[Dataset]:
+    """
+    Take jelly file and return generators based on the detected logical type.
 
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_QUADS:
-        yield parse_flat_quads_stream(frames=frames, options=options, store=store)
-        return
+    Yields one graph/dataset per frame.
 
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_GRAPHS:
-        yield from parse_graph_stream(frames=frames, options=options, store=store)
-        return
+    Args:
+        inp (IO[bytes]): input jelly buffered binary stream
+        graph_factory (Callable): lambda to construct a Graph
+        dataset_factory (Callable): lambda to construct a Dataset
 
-    logical_type_name = jelly.LogicalStreamType.Name(options.stream_types.logical_type)
-    msg = f"the stream type {logical_type_name} is not supported "
-    raise NotImplementedError(msg)
+    Raises:
+        NotImplementedError: is raised if a logical type is not implemented
 
+    Yields:
+        Generator[Any] | Generator[Dataset] | Generator[Graph]:
+            returns generators for graphs/datasets based on the type of input
 
-def graph_from_jelly(
-    inp: IO[bytes],
-    store: Store | str = "default",
-    identifier: str | None = None,
-) -> Any | Dataset | Graph:
+    """
     options, frames = get_options_and_frames(inp)
 
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_DATASETS:
-        msg = (
-            "the stream contains multiple datasets and cannot be parsed into "
-            "a single dataset"
-        )
-        raise NotImplementedError(msg)
-
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_TRIPLES:
-        return parse_flat_triples_stream(
+    if options.stream_types.physical_type == jelly.PHYSICAL_STREAM_TYPE_TRIPLES:
+        yield from parse_triples_stream(
             frames=frames,
             options=options,
-            store=store,
-            identifier=identifier,
+            graph_factory=graph_factory,
+            parsing_mode=ParsingMode.GROUPED,
         )
+        return
 
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_FLAT_QUADS:
-        return parse_flat_quads_stream(
+    if options.stream_types.physical_type in (
+        jelly.PHYSICAL_STREAM_TYPE_QUADS,
+        jelly.PHYSICAL_STREAM_TYPE_GRAPHS,
+    ):
+        yield from parse_quads_stream(
             frames=frames,
             options=options,
-            store=store,
-            identifier=identifier,
+            dataset_factory=dataset_factory,
+            parsing_mode=ParsingMode.GROUPED,
         )
+        return
 
-    if options.stream_types.logical_type == jelly.LOGICAL_STREAM_TYPE_GRAPHS:
-        ds = Dataset(store=store, default_union=True)
-        ds.default_context = Graph(identifier=identifier, store=store)
+    physical_type_name = jelly.PhysicalStreamType.Name(
+        options.stream_types.physical_type
+    )
+    msg = f"the stream type {physical_type_name} is not supported "
+    raise NotImplementedError(msg)
 
-        for graph in parse_graph_stream(frames=frames, options=options, store=store):
-            ds.add_graph(graph)
 
-        return ds
+def parse_jelly_flat(
+    inp: IO[bytes],
+    graph_factory: Callable[[], Graph],
+    dataset_factory: Callable[[], Dataset],
+) -> Any | Dataset | Graph:
+    """
+    Parse jelly file with FLAT physical type into one Graph/Dataset.
 
-    logical_type_name = jelly.LogicalStreamType.Name(options.stream_types.logical_type)
-    msg = f"the stream type {logical_type_name} is not supported "
+    Args:
+        inp (IO[bytes]): input jelly buffered binary stream
+        graph_factory (Callable): lambda to construct a Graph
+        dataset_factory (Callable): lambda to construct a Dataset
+
+    Raises:
+        NotImplementedError: if physical type is not supported
+
+    Returns:
+        RDFLib Graph or Dataset
+
+    """
+    options, frames = get_options_and_frames(inp)
+
+    if options.stream_types.physical_type == jelly.PHYSICAL_STREAM_TYPE_TRIPLES:
+        return next(
+            parse_triples_stream(
+                frames=frames,
+                options=options,
+                graph_factory=graph_factory,
+                parsing_mode=ParsingMode.FLAT,
+            )
+        )
+
+    if options.stream_types.physical_type in (
+        jelly.PHYSICAL_STREAM_TYPE_QUADS,
+        jelly.PHYSICAL_STREAM_TYPE_GRAPHS,
+    ):
+        return next(
+            parse_quads_stream(
+                frames=frames,
+                options=options,
+                dataset_factory=dataset_factory,
+                parsing_mode=ParsingMode.FLAT,
+            )
+        )
+    physical_type_name = jelly.PhysicalStreamType.Name(
+        options.stream_types.physical_type
+    )
+    msg = f"the stream type {physical_type_name} is not supported "
     raise NotImplementedError(msg)
 
 
 class RDFLibJellyParser(RDFLibParser):
     def parse(self, source: InputSource, sink: Graph) -> None:
+        """
+        Parse jelly file into provided RDFLib Graph.
+
+        Args:
+            source (InputSource): jelly file as buffered binary stream InputSource obj
+            sink (Graph): RDFLib Graph
+
+        Raises:
+            TypeError: raises error if invalid input
+
+        """
         inp = source.getByteStream()  # type: ignore[no-untyped-call]
         if inp is None:
             msg = "expected source to be a stream of bytes"
             raise TypeError(msg)
-        graph_from_jelly(
+        parse_jelly_flat(
             inp,
-            identifier=sink.identifier,
-            store=sink.store,
+            graph_factory=lambda: Graph(store=sink.store, identifier=sink.identifier),
+            dataset_factory=lambda: Dataset(store=sink.store),
         )

pyjelly/integrations/rdflib/serialize.py

@@ -23,6 +23,17 @@ from pyjelly.serialize.streams import (
 
 class RDFLibTermEncoder(TermEncoder):
     def encode_any(self, term: object, slot: Slot) -> RowsAndTerm:
+        """
+        Encode term based on its RDFLib object.
+
+        Args:
+            term (object): term to encode
+            slot (Slot): its place in statement.
+
+        Returns:
+            RowsAndTerm: encoded extra rows and a jelly term to encode
+
+        """
         if slot is Slot.graph and term == DATASET_DEFAULT_GRAPH_ID:
             return self.encode_default_graph()
 
@@ -60,6 +71,21 @@ def triples_stream_frames(
     stream: TripleStream,
     data: Graph | Dataset,
 ) -> Generator[jelly.RdfStreamFrame]:
+    """
+    Serialize a Graph/Dataset into jelly frames.
+
+    Args:
+        stream (TripleStream): stream that specifies triples processing
+        data (Graph | Dataset): Graph/Dataset to serialize.
+
+    Notes:
+        if Dataset is given, its graphs are unpacked and iterated over
+        if flow is GraphsFrameFlow, emits a frame per graph.
+
+    Yields:
+        Generator[jelly.RdfStreamFrame]: jelly frames.
+
+    """
     stream.enroll()
     if stream.options.params.namespace_declarations:
         namespace_declarations(data, stream)
@@ -68,6 +94,7 @@ def triples_stream_frames(
         for terms in graph:
             if frame := stream.triple(terms):
                 yield frame
+        # this part turns each graph to a frame for graphs logical type
         if frame := stream.flow.frame_from_graph():
             yield frame
     if stream.stream_types.flat and (frame := stream.flow.to_stream_frame()):
@@ -79,6 +106,20 @@ def quads_stream_frames(
     stream: QuadStream,
     data: Dataset,
 ) -> Generator[jelly.RdfStreamFrame]:
+    """
+    Serialize a Dataset into jelly frames.
+
+    Notes:
+        Emits one frame per dataset if flow is of DatasetsFrameFlow.
+
+    Args:
+        stream (QuadStream): stream that specifies quads processing
+        data (Dataset): Dataset to serialize.
+
+    Yields:
+        Generator[jelly.RdfStreamFrame]: jelly frames
+
+    """
     assert isinstance(data, Dataset)
     stream.enroll()
     if stream.options.params.namespace_declarations:
@@ -97,6 +138,21 @@ def graphs_stream_frames(
     stream: GraphStream,
     data: Dataset,
 ) -> Generator[jelly.RdfStreamFrame]:
+    """
+    Serialize a Dataset into jelly frames as a stream of graphs.
+
+    Notes:
+        If flow of DatasetsFrameFlow type, the whole dataset
+        will be encoded into one frame.
+
+    Args:
+        stream (GraphStream): stream that specifies graphs processing
+        data (Dataset): Dataset to serialize.
+
+    Yields:
+        Generator[jelly.RdfStreamFrame]: jelly frames
+
+    """
     assert isinstance(data, Dataset)
     stream.enroll()
     if stream.options.params.namespace_declarations:
@@ -171,6 +227,18 @@ class RDFLibJellySerializer(RDFLibSerializer):
         options: SerializerOptions | None = None,
         **unused: Any,
     ) -> None:
+        """
+        Serialize self.store content to Jelly format.
+
+        Args:
+            out (IO[bytes]): output buffered writer
+            stream (Stream | None, optional): Jelly stream object. Defaults to None.
+            options (SerializerOptions | None, optional): Serializer options
+                if defined beforehand, e.g., read from a separate file.
+                Defaults to None.
+            **unused(Any): unused args for RDFLib serialize
+
+        """
         if options is None:
             options = self.guess_options()
         if stream is None: