fix serialiser makes invalid turtle curis

rdflib/plugins/serializers/longturtle.py

@@ -18,6 +18,7 @@
 
 from __future__ import annotations
 
+import warnings
 from typing import IO, Any, Optional
 
 from rdflib.compare import to_canonical_graph
@@ -166,14 +167,14 @@ def preprocessTriple(self, triple: _TripleType) -> None:
                     # predicate corresponds to base namespace
                     continue
             # Don't use generated prefixes for subjects and objects
-            self.getQName(node, gen_prefix=(i == VERB))
+            self.get_pname(node, gen_prefix=(i == VERB))
             if isinstance(node, Literal) and node.datatype:
-                self.getQName(node.datatype, gen_prefix=_GEN_QNAME_FOR_DT)
+                self.get_pname(node.datatype, gen_prefix=_GEN_QNAME_FOR_DT)
         p = triple[1]
         if isinstance(p, BNode):  # hmm - when is P ever a bnode?
             self._references[p] += 1
 
-    def getQName(self, uri, gen_prefix=True):
+    def get_pname(self, uri, gen_prefix=True):
         if not isinstance(uri, URIRef):
             return None
 
@@ -191,16 +192,30 @@ def getQName(self, uri, gen_prefix=True):
 
         prefix, namespace, local = parts
 
+        # To understand treatment of % character refer to Productions for terminal PLX at
+        # https://www.w3.org/TR/turtle/#grammar-production-PLX
+        # Only % NOT followed by two hex chars requires manual backslash escaping
         local = local.replace(r"(", r"\(").replace(r")", r"\)")
+        local = self.LOCALNAME_PECRENT_CHARACTER_REQUIRING_ESCAPE_REGEX.sub(
+            "\\%", local
+        )
 
-        # QName cannot end with .
+        # PName cannot end with .
         if local.endswith("."):
             return None
 
         prefix = self.addNamespace(prefix, namespace)
 
         return "%s:%s" % (prefix, local)
 
+    def getQName(self, uri, gen_prefix=True):
+        warnings.warn(
+            "LongTurtleSerializer.getQName is deprecated, use LongTurtleSerializer.get_pname instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.get_pname(uri, gen_prefix)
+
     def startDocument(self):
         self._started = True
         ns_list = sorted(self.namespaces.items())
@@ -256,12 +271,12 @@ def label(self, node, position):
         if isinstance(node, Literal):
             return node._literal_n3(
                 use_plain=True,
-                qname_callback=lambda dt: self.getQName(dt, _GEN_QNAME_FOR_DT),
+                qname_callback=lambda dt: self.get_pname(dt, _GEN_QNAME_FOR_DT),
             )
         else:
             node = self.relativize(node)
 
-            return self.getQName(node, position == VERB) or node.n3()
+            return self.get_pname(node, position == VERB) or node.n3()
 
     def p_squared(
         self,

rdflib/plugins/serializers/n3.py

@@ -2,6 +2,8 @@
 Notation 3 (N3) RDF graph serializer for RDFLib.
 """
 
+import warnings
+
 from rdflib.graph import Graph
 from rdflib.namespace import OWL, Namespace
 from rdflib.plugins.serializers.turtle import OBJECT, SUBJECT, TurtleSerializer
@@ -47,14 +49,22 @@ def preprocessTriple(self, triple):  # noqa: N802
             for t in triple[2]:
                 self.preprocessTriple(t)
 
-    def getQName(self, uri, gen_prefix=True):  # noqa: N802
+    def get_pname(self, uri, gen_prefix=True):
         qname = None
         if self.parent is not None:
-            qname = self.parent.getQName(uri, gen_prefix)
+            qname = self.parent.get_pname(uri, gen_prefix)
         if qname is None:
-            qname = super(N3Serializer, self).getQName(uri, gen_prefix)
+            qname = super(N3Serializer, self).get_pname(uri, gen_prefix)
         return qname
 
+    def getQName(self, uri, gen_prefix=True):  # noqa: N802
+        warnings.warn(
+            "N3Serializer.getQName is deprecated, use N3Serializer.get_pname instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.get_pname(uri, gen_prefix)
+
     def statement(self, subject):
         self.subjectDone(subject)
         properties = self.buildPredicateHash(subject)

rdflib/plugins/serializers/trig.py

@@ -45,7 +45,7 @@ def preprocess(self) -> None:
                 continue
             self.store = context
             # Don't generate a new prefix for a graph URI if one already exists
-            self.getQName(context.identifier, False)
+            self.get_pname(context.identifier, False)
             self._subjects = {}
 
             for triple in context:
@@ -103,7 +103,7 @@ def serialize(
                     iri = store.identifier.n3()
                 else:
                     # Show the full graph URI if a prefix for it doesn't already exist
-                    iri = self.getQName(store.identifier, False)
+                    iri = self.get_pname(store.identifier, False)
                     if iri is None:
                         iri = store.identifier.n3()
                 self.write(self.indent() + "\n%s {" % iri)

rdflib/plugins/serializers/turtle.py

@@ -5,6 +5,8 @@
 
 from __future__ import annotations
 
+import re
+import warnings
 from collections import defaultdict
 from typing import (
     IO,
@@ -43,6 +45,9 @@ class RecursiveSerializer(Serializer):
     maxDepth = 10
     indentString = "  "
     roundtrip_prefixes: Tuple[Any, ...] = ()
+    LOCALNAME_PECRENT_CHARACTER_REQUIRING_ESCAPE_REGEX = re.compile(
+        r"%(?![0-9A-Fa-f]{2})"
+    )
 
     def __init__(self, store: Graph):
         super(RecursiveSerializer, self).__init__(store)
@@ -204,6 +209,9 @@ class TurtleSerializer(RecursiveSerializer):
 
     short_name = "turtle"
     indentString = "    "
+    LOCALNAME_PECRENT_CHARACTER_REQUIRING_ESCAPE_REGEX = re.compile(
+        r"%(?![0-9A-Fa-f]{2})"
+    )
 
     def __init__(self, store: Graph):
         self._ns_rewrite: Dict[str, str] = {}
@@ -300,15 +308,15 @@ def preprocessTriple(self, triple: _TripleType) -> None:
                     # predicate corresponds to base namespace
                     continue
             # Don't use generated prefixes for subjects and objects
-            self.getQName(node, gen_prefix=(i == VERB))
+            self.get_pname(node, gen_prefix=(i == VERB))
             if isinstance(node, Literal) and node.datatype:
-                self.getQName(node.datatype, gen_prefix=_GEN_QNAME_FOR_DT)
+                self.get_pname(node.datatype, gen_prefix=_GEN_QNAME_FOR_DT)
         p = triple[1]
         if isinstance(p, BNode):  # hmm - when is P ever a bnode?
             self._references[p] += 1
 
-    # TODO: Rename to get_pname
-    def getQName(self, uri: Node, gen_prefix: bool = True) -> Optional[str]:
+    # Refer to Productions for terminals PNAME_NS and PNAME_LN https://www.w3.org/TR/turtle/#sec-grammar-grammar
+    def get_pname(self, uri: Node, gen_prefix: bool = True) -> Optional[str]:
         if not isinstance(uri, URIRef):
             return None
 
@@ -328,16 +336,30 @@ def getQName(self, uri: Node, gen_prefix: bool = True) -> Optional[str]:
 
         prefix, namespace, local = parts
 
+        # To understand treatment of % character refer to Productions for terminal PLX at
+        # https://www.w3.org/TR/turtle/#grammar-production-PLX
+        # Only % NOT followed by two hex chars requires manual backslash escaping
         local = local.replace(r"(", r"\(").replace(r")", r"\)")
+        local = self.LOCALNAME_PECRENT_CHARACTER_REQUIRING_ESCAPE_REGEX.sub(
+            "\\%", local
+        )
 
-        # QName cannot end with .
+        # PName cannot end with .
         if local.endswith("."):
             return None
 
         prefix = self.addNamespace(prefix, namespace)
 
         return "%s:%s" % (prefix, local)
 
+    def getQName(self, uri: Node, gen_prefix: bool = True) -> Optional[str]:
+        warnings.warn(
+            "TurtleSerializer.getQName is deprecated, use TurtleSerializer.get_pname instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.get_pname(uri, gen_prefix)
+
     def startDocument(self) -> None:
         self._started = True
         ns_list = sorted(self.namespaces.items())
@@ -393,12 +415,12 @@ def label(self, node: Node, position: int) -> str:
         if isinstance(node, Literal):
             return node._literal_n3(
                 use_plain=True,
-                qname_callback=lambda dt: self.getQName(dt, _GEN_QNAME_FOR_DT),
+                qname_callback=lambda dt: self.get_pname(dt, _GEN_QNAME_FOR_DT),
             )
         else:
             node = self.relativize(node)  # type: ignore[type-var]
 
-            return self.getQName(node, position == VERB) or node.n3()
+            return self.get_pname(node, position == VERB) or node.n3()
 
     def p_squared(self, node: Node, position: int, newline: bool = False) -> bool:
         if (

test/data/roundtrip/iri_with_escaped_percent.ttl

@@ -0,0 +1,4 @@
+@prefix : <#> .
+:zzz_\%_zzz :prop "test iri including backslash-escaped percent char \\%" .
+:zzz%20zzz :prop "test iri including percent-encoded space char %20" .
+:zzz\%azzz :prop "test iri including backslash-escpaed percent char followed by single hex character" .

test/test_deprecation_notice.py

@@ -0,0 +1,72 @@
+import pytest
+
+from rdflib import Dataset, Graph, Namespace, URIRef
+from rdflib.plugins.serializers.longturtle import LongTurtleSerializer
+from rdflib.plugins.serializers.n3 import N3Serializer
+from rdflib.plugins.serializers.turtle import TurtleSerializer
+
+
+def test_dataset_contexts_method():
+    ds = Dataset()
+    with pytest.warns(
+        DeprecationWarning,
+        match="Dataset.contexts is deprecated, use Dataset.graphs instead.",
+    ):
+        # Call list() to consume the generator to emit the warning.
+        list(ds.contexts())
+
+
+def test_dataset_default_context_property():
+    ds = Dataset()
+    with pytest.warns(
+        DeprecationWarning,
+        match="Dataset.default_context is deprecated, use Dataset.default_graph instead.",
+    ):
+        ds.default_context
+
+    with pytest.warns(
+        DeprecationWarning,
+        match="Dataset.default_context is deprecated, use Dataset.default_graph instead.",
+    ):
+        ds.default_context = ds.graph()
+
+
+def test_dataset_identifier_property():
+    ds = Dataset()
+    with pytest.warns(
+        DeprecationWarning,
+        match="Dataset.identifier is deprecated and will be removed in future versions.",
+    ):
+        ds.identifier
+
+
+@pytest.mark.parametrize(
+    ("serializer_cls", "warning_message"),
+    [
+        (
+            TurtleSerializer,
+            "TurtleSerializer.getQName is deprecated, use TurtleSerializer.get_pname instead.",
+        ),
+        (
+            LongTurtleSerializer,
+            "LongTurtleSerializer.getQName is deprecated, use LongTurtleSerializer.get_pname instead.",
+        ),
+        (
+            N3Serializer,
+            "N3Serializer.getQName is deprecated, use N3Serializer.get_pname instead.",
+        ),
+    ],
+)
+def test_serializer_getqname_method(
+    serializer_cls,
+    warning_message: str,
+):
+    graph = Graph()
+    ex = Namespace("http://example.org/")
+    graph.bind("ex", ex)
+    serializer = serializer_cls(graph)
+
+    with pytest.warns(DeprecationWarning, match=warning_message):
+        qname = serializer.getQName(URIRef("http://example.org/value"))
+
+    assert qname == "ex:value"

test/test_roundtrip.py

@@ -526,6 +526,7 @@ def test_n3_suite(
     (TEST_DATA_DIR / "variants" / "diverse_quads.nq", "nquads"),
     (TEST_DATA_DIR / "variants" / "diverse_quads.trig", "trig"),
     (TEST_DATA_DIR / "roundtrip" / "bnode_refs.trig", "trig"),
+    (TEST_DATA_DIR / "roundtrip" / "iri_with_escaped_percent.ttl", "ttl"),
     (TEST_DATA_DIR / "example-lots_of_graphs.n3", "n3"),
     (TEST_DATA_DIR / "issue156.n3", "n3"),
 ]

test/test_trig.py

@@ -73,7 +73,7 @@ def test_graph_qname_syntax():
 
 def test_graph_uri_syntax():
     g = rdflib.Dataset()
-    # getQName will not abbreviate this, so it should serialize as
+    # get_pname will not abbreviate this, so it should serialize as
     # a '<...>' term.
     g.add(TRIPLE + (rdflib.URIRef("http://example.com/foo."),))
     out = g.serialize(format="trig", encoding="latin-1")