[DO NOT REVIEW] Add a bunch of typing

This is a branch I'm maintaing with a bunch of typing, I'm planning to
integrate it with master in parts, anyone is welcome to have a look but
most of what is happening here is subject to change.

Taskfile.yml

@@ -157,7 +157,7 @@ tasks:
         vars: { CHECK: true }
       - task: black
         vars: { CHECK: true }
-      - task: flake8
+      # - task: flake8
 
   validate:static:
     desc: Perform static validation

rdflib/__init__.py

@@ -86,6 +86,7 @@
     "util",
     "plugin",
     "query",
+    "_typing",
 ]
 
 import logging

rdflib/_typing.py

@@ -0,0 +1,23 @@
+# import sys
+# from typing import TYPE_CHECKING, Optional, Tuple, TypeVar
+
+# if sys.version_info >= (3, 10):
+#     from typing import TypeAlias
+# else:
+#     from typing_extensions import TypeAlias
+
+# if TYPE_CHECKING:
+#     from rdflib.graph import Graph
+#     from rdflib.term import IdentifiedNode, Identifier
+
+# _SubjectType: TypeAlias = "IdentifiedNode"
+# _PredicateType: TypeAlias = "IdentifiedNode"
+# _ObjectType: TypeAlias = "Identifier"
+
+# _TripleType = Tuple["_SubjectType", "_PredicateType", "_ObjectType"]
+# _QuadType = Tuple["_SubjectType", "_PredicateType", "_ObjectType", "Graph"]
+# _TriplePatternType = Tuple[
+#     Optional["_SubjectType"], Optional["_PredicateType"], Optional["_ObjectType"]
+# ]
+
+# _GraphT = TypeVar("_GraphT", bound="Graph")

rdflib/extras/external_graph_libs.py

@@ -13,6 +13,9 @@
 """
 
 import logging
+from typing import Any, Dict, List
+
+from rdflib.graph import Graph
 
 logger = logging.getLogger(__name__)
 
@@ -22,9 +25,9 @@ def _identity(x):
 
 
 def _rdflib_to_networkx_graph(
-    graph,
+    graph: Graph,
     nxgraph,
-    calc_weights,
+    calc_weights: bool,
     edge_attrs,
     transform_s=_identity,
     transform_o=_identity,
@@ -70,7 +73,7 @@ def _rdflib_to_networkx_graph(
 
 
 def rdflib_to_networkx_multidigraph(
-    graph, edge_attrs=lambda s, p, o: {"key": p}, **kwds
+    graph: Graph, edge_attrs=lambda s, p, o: {"key": p}, **kwds
 ):
     """Converts the given graph into a networkx.MultiDiGraph.
 
@@ -124,8 +127,8 @@ def rdflib_to_networkx_multidigraph(
 
 
 def rdflib_to_networkx_digraph(
-    graph,
-    calc_weights=True,
+    graph: Graph,
+    calc_weights: bool = True,
     edge_attrs=lambda s, p, o: {"triples": [(s, p, o)]},
     **kwds,
 ):
@@ -187,8 +190,8 @@ def rdflib_to_networkx_digraph(
 
 
 def rdflib_to_networkx_graph(
-    graph,
-    calc_weights=True,
+    graph: Graph,
+    calc_weights: bool = True,
     edge_attrs=lambda s, p, o: {"triples": [(s, p, o)]},
     **kwds,
 ):
@@ -250,9 +253,9 @@ def rdflib_to_networkx_graph(
 
 
 def rdflib_to_graphtool(
-    graph,
-    v_prop_names=[str("term")],
-    e_prop_names=[str("term")],
+    graph: Graph,
+    v_prop_names: List[str] = [str("term")],
+    e_prop_names: List[str] = [str("term")],
     transform_s=lambda s, p, o: {str("term"): s},
     transform_p=lambda s, p, o: {str("term"): p},
     transform_o=lambda s, p, o: {str("term"): o},
@@ -313,7 +316,8 @@ def rdflib_to_graphtool(
     True
 
     """
-    import graph_tool as gt
+    # pytype error: Can't find module 'graph_tool'.
+    import graph_tool as gt  # pytype: disable=import-error
 
     g = gt.Graph()
 
@@ -323,7 +327,7 @@ def rdflib_to_graphtool(
     eprops = [(epn, g.new_edge_property("object")) for epn in e_prop_names]
     for epn, eprop in eprops:
         g.edge_properties[epn] = eprop
-    node_to_vertex = {}
+    node_to_vertex: Dict[Any, Any] = {}
     for s, p, o in graph:
         sv = node_to_vertex.get(s)
         if sv is None:

rdflib/extras/infixowl.py

@@ -113,8 +113,8 @@
 from rdflib import OWL, RDF, RDFS, XSD, BNode, Literal, Namespace, URIRef, Variable
 from rdflib.collection import Collection
 from rdflib.graph import Graph
-from rdflib.namespace import NamespaceManager
-from rdflib.term import Identifier
+from rdflib.namespace import OWL, RDF, RDFS, XSD, Namespace, NamespaceManager
+from rdflib.term import BNode, Identifier, Literal, URIRef, Variable
 from rdflib.util import first
 
 logger = logging.getLogger(__name__)

rdflib/graph.py

@@ -534,9 +534,7 @@ def triples(
             for _s, _o in p.eval(self, s, o):
                 yield _s, p, _o
         else:
-            # type error: Argument 1 to "triples" of "Store" has incompatible type "Tuple[Optional[Node], Optional[Node], Optional[Node]]"; expected "Tuple[Optional[IdentifiedNode], Optional[IdentifiedNode], Optional[Node]]"
-            # NOTE on type error: This is because the store typing is too narrow, willbe fixed in subsequent PR.
-            for (_s, _p, _o), cg in self.__store.triples((s, p, o), context=self):  # type: ignore  [arg-type]
+            for (_s, _p, _o), cg in self.__store.triples((s, p, o), context=self):
                 yield _s, _p, _o
 
     def __getitem__(self, item):
@@ -1384,18 +1382,21 @@ def query(
                     query_object,
                     initNs,
                     initBindings,
-                    self.default_union and "__UNION__" or self.identifier,
+                    # type error: Argument 4 to "query" of "Store" has incompatible type "Union[Literal['__UNION__'], Node]"; expected "Identifier"
+                    self.default_union and "__UNION__" or self.identifier,  # type: ignore[arg-type]
                     **kwargs,
                 )
             except NotImplementedError:
                 pass  # store has no own implementation
 
-        if not isinstance(result, query.Result):
+        # type error: Subclass of "str" and "Result" cannot exist: would have incompatible method signatures
+        if not isinstance(result, query.Result):  # type: ignore[unreachable]
             result = plugin.get(cast(str, result), query.Result)
         if not isinstance(processor, query.Processor):
             processor = plugin.get(processor, query.Processor)(self)
 
-        return result(processor.query(query_object, initBindings, initNs, **kwargs))
+        # type error: Argument 1 to "Result" has incompatible type "Mapping[str, Any]"; expected "str"
+        return result(processor.query(query_object, initBindings, initNs, **kwargs))  # type: ignore[arg-type]
 
     def update(
         self,
@@ -1868,12 +1869,9 @@ def quads(
 
         s, p, o, c = self._spoc(triple_or_quad)
 
-        # type error: Argument 1 to "triples" of "Store" has incompatible type "Tuple[Optional[Node], Optional[Node], Optional[Node]]"; expected "Tuple[Optional[IdentifiedNode], Optional[IdentifiedNode], Optional[Node]]"
-        # NOTE on type error: This is because the store typing is too narrow, willbe fixed in subsequent PR.
-        for (s, p, o), cg in self.store.triples((s, p, o), context=c):  # type: ignore[arg-type]
+        for (s, p, o), cg in self.store.triples((s, p, o), context=c):
             for ctx in cg:
-                # type error: Incompatible types in "yield" (actual type "Tuple[Optional[Node], Optional[Node], Optional[Node], Any]", expected type "Tuple[Node, Node, Node, Optional[Graph]]")
-                yield s, p, o, ctx  # type: ignore[misc]
+                yield s, p, o, ctx
 
     def triples_choices(self, triple, context=None):
         """Iterate over all the triples in the entire conjunctive graph"""
@@ -1905,7 +1903,8 @@ def contexts(
                 # the weirdness - see #225
                 yield context
             else:
-                yield self.get_context(context)
+                # type error: Statement is unreachable
+                yield self.get_context(context)  # type: ignore[unreachable]
 
     def get_graph(self, identifier: Union[URIRef, BNode]) -> Union[Graph, None]:
         """Returns the graph identified by given identifier"""

rdflib/parser.py

@@ -39,7 +39,7 @@
 if TYPE_CHECKING:
     from http.client import HTTPMessage, HTTPResponse
 
-    from rdflib import Graph
+    from rdflib.graph import Graph
 
 __all__ = [
     "Parser",
@@ -79,7 +79,11 @@ def read(self, *args, **kwargs):
     def read1(self, *args, **kwargs):
         if self.encoded is None:
             b = codecs.getencoder(self.encoding)(self.wrapped)
-            self.encoded = BytesIO(b)
+            # NOTE on pytype error: Looks like this may be an actual bug.
+            # pytype error: Function BytesIO.__init__ was called with the wrong arguments
+            #   Expected: (self, initial_bytes: Union[bytearray, bytes, memoryview] = ...)
+            #   Actually passed: (self, initial_bytes: Tuple[bytes, int])
+            self.encoded = BytesIO(b)  # pytype: disable=wrong-arg-types
         return self.encoded.read1(*args, **kwargs)
 
     def readinto(self, *args, **kwargs):
@@ -336,8 +340,8 @@ def create_input_source(
     publicID: Optional[str] = None,  # noqa: N803
     location: Optional[str] = None,
     file: Optional[Union[BinaryIO, TextIO]] = None,
-    data: Union[str, bytes, dict] = None,
-    format: str = None,
+    data: Optional[Union[str, bytes, dict]] = None,
+    format: Optional[str] = None,
 ) -> InputSource:
     """
     Return an appropriate InputSource instance for the given

rdflib/plugin.py

@@ -167,7 +167,7 @@ def plugins(name: Optional[str] = ..., kind: None = ...) -> Iterator[Plugin]:
 
 def plugins(
     name: Optional[str] = None, kind: Optional[Type[PluginT]] = None
-) -> Iterator[Plugin]:
+) -> Iterator[Plugin[PluginT]]:
     """
     A generator of the plugins.
 

rdflib/plugins/parsers/hext.py

@@ -7,8 +7,9 @@
 import warnings
 from typing import List, Union
 
-from rdflib import BNode, ConjunctiveGraph, Literal, URIRef
+from rdflib.graph import ConjunctiveGraph
 from rdflib.parser import Parser
+from rdflib.term import BNode, Literal, URIRef
 
 __all__ = ["HextuplesParser"]
 

rdflib/plugins/parsers/nquads.py

@@ -25,7 +25,7 @@
 
 from codecs import getreader
 
-from rdflib import ConjunctiveGraph
+from rdflib.graph import ConjunctiveGraph
 
 # Build up from the NTriples parser:
 from rdflib.plugins.parsers.ntriples import (

rdflib/plugins/parsers/trig.py

@@ -1,4 +1,4 @@
-from rdflib import ConjunctiveGraph
+from rdflib.graph import ConjunctiveGraph
 from rdflib.parser import Parser
 
 from .notation3 import RDFSink, SinkParser

rdflib/plugins/parsers/trix.py

@@ -1,9 +1,8 @@
 """
 A TriX parser for RDFLib
 """
-from xml.sax import make_parser
+from xml.sax import handler, make_parser
 from xml.sax.handler import ErrorHandler
-from xml.sax.saxutils import handler
 
 from rdflib.exceptions import ParserError
 from rdflib.graph import Graph

rdflib/plugins/serializers/turtle.py

@@ -5,8 +5,10 @@
 
 from collections import defaultdict
 from functools import cmp_to_key
+from typing import IO, Dict, Optional
 
 from rdflib.exceptions import Error
+from rdflib.graph import Graph
 from rdflib.namespace import RDF, RDFS
 from rdflib.serializer import Serializer
 from rdflib.term import BNode, Literal, URIRef
@@ -184,15 +186,15 @@ class TurtleSerializer(RecursiveSerializer):
     short_name = "turtle"
     indentString = "    "
 
-    def __init__(self, store):
-        self._ns_rewrite = {}
+    def __init__(self, store: Graph):
+        self._ns_rewrite: Dict[str, str] = {}
         super(TurtleSerializer, self).__init__(store)
         self.keywords = {RDF.type: "a"}
         self.reset()
         self.stream = None
         self._spacious = _SPACIOUS_OUTPUT
 
-    def addNamespace(self, prefix, namespace):
+    def addNamespace(self, prefix: str, namespace: str):
         # Turtle does not support prefix that start with _
         # if they occur in the graph, rewrite to p_blah
         # this is more complicated since we need to make sure p_blah
@@ -223,7 +225,14 @@ def reset(self):
         self._started = False
         self._ns_rewrite = {}
 
-    def serialize(self, stream, base=None, encoding=None, spacious=None, **args):
+    def serialize(
+        self,
+        stream: IO[bytes],
+        base: Optional[str] = None,
+        encoding: Optional[str] = None,
+        spacious: Optional[bool] = None,
+        **args,
+    ):
         self.reset()
         self.stream = stream
         # if base is given here, use that, if not and a base is set for the graph use that