prov 2.5.2__tar.gz → 2.5.4__tar.gz

{prov-2.5.2/src/prov.egg-info → prov-2.5.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: prov
-Version: 2.5.2
+Version: 2.5.4
 Summary: A library for W3C Provenance Data Model supporting PROV-JSON, PROV-XML and PROV-O (RDF)
 Author-email: Trung Dong Huynh <trungdong@donggiang.com>
 License-Expression: MIT
@@ -21,6 +21,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: 3.14
+Classifier: Programming Language :: Python :: 3.15
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy
@@ -36,7 +37,7 @@ Requires-Dist: python-dateutil>=2.2
 Provides-Extra: rdf
 Requires-Dist: rdflib<8,>=6.0.0; extra == "rdf"
 Provides-Extra: xml
-Requires-Dist: lxml>=3.3.5; extra == "xml"
+Requires-Dist: lxml>=6.1.3; extra == "xml"
 Provides-Extra: plot
 Requires-Dist: matplotlib>=3.6; extra == "plot"
 Dynamic: license-file
@@ -81,7 +82,7 @@ Feedback is welcome on the [issue tracker](https://github.com/trungdong/prov/iss
 ## Supported versions
 
 The latest 3.x release receives all fixes. This 2.x line receives security
-fixes, plus bug fixes back-ported from 3.x up to and including 2.6.0, after
-which it reverts to security fixes only; 1.x and earlier no longer receive
-fixes. See [SECURITY.md](https://github.com/trungdong/prov/blob/2.x/SECURITY.md)
+fixes only; 2.5.4 was the last release to carry bug fixes back-ported from
+3.x. 1.x and earlier no longer receive fixes. See
+[SECURITY.md](https://github.com/trungdong/prov/blob/2.x/SECURITY.md)
 for the full support table and how to report a vulnerability.

{prov-2.5.2 → prov-2.5.4}/README.md

@@ -38,7 +38,7 @@ Feedback is welcome on the [issue tracker](https://github.com/trungdong/prov/iss
 ## Supported versions
 
 The latest 3.x release receives all fixes. This 2.x line receives security
-fixes, plus bug fixes back-ported from 3.x up to and including 2.6.0, after
-which it reverts to security fixes only; 1.x and earlier no longer receive
-fixes. See [SECURITY.md](https://github.com/trungdong/prov/blob/2.x/SECURITY.md)
+fixes only; 2.5.4 was the last release to carry bug fixes back-ported from
+3.x. 1.x and earlier no longer receive fixes. See
+[SECURITY.md](https://github.com/trungdong/prov/blob/2.x/SECURITY.md)
 for the full support table and how to report a vulnerability.

{prov-2.5.2 → prov-2.5.4}/pyproject.toml

@@ -21,6 +21,7 @@ classifiers = [
     "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3.13",
     "Programming Language :: Python :: 3.14",
+    "Programming Language :: Python :: 3.15",
     "Programming Language :: Python :: 3 :: Only",
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
@@ -41,7 +42,7 @@ rdf = [
     "rdflib>=6.0.0,<8",
 ]
 xml = [
-    "lxml>=3.3.5",
+    "lxml>=6.1.3",
 ]
 plot = [
     "matplotlib>=3.6",
@@ -59,7 +60,7 @@ Issues = "https://github.com/trungdong/prov/issues"
 Changelog = "https://prov.readthedocs.io/en/latest/history.html"
 
 [build-system]
-requires = ["setuptools>=40.8.0", "wheel"]  # PEP 508 specifications.
+requires = ["setuptools>=84.2.1", "wheel"]  # PEP 508 specifications.
 build-backend = "setuptools.build_meta"
 
 [tool.setuptools.dynamic]
@@ -80,17 +81,17 @@ prov = ["py.typed"]
 
 [dependency-groups]
 dev = [
-    "coverage>=7.6.10",
-    "hypothesis>=6.156.1",
+    "coverage>=7.16.2",
+    "hypothesis>=6.168.5",
     "jsonschema>=4",
     "lxml-stubs>=0.5.1",
-    "mypy>=1.19.1",
-    "pre-commit>=4.0.1",
+    "mypy>=2.4.0",
+    "pre-commit>=4.6.2",
     "pytest>=8.4.2",
     "pytest-cov>=7.1.0",
-    "ruff>=0.15.20",
-    "types-networkx>=3.4.2.20250509",
-    "types-python-dateutil>=2.9.0.20260124",
+    "ruff>=0.17.3",
+    "types-networkx>=3.6.1.20261029",
+    "types-python-dateutil>=2.9.0.20261029",
 ]
 docs = [
     "sphinx>=8.1.3,<9",

{prov-2.5.2 → prov-2.5.4}/src/prov/__init__.py

@@ -11,7 +11,7 @@ if TYPE_CHECKING:
 
 __author__ = "Trung Dong Huynh"
 __email__ = "trungdong@donggiang.com"
-__version__ = "2.5.2"
+__version__ = "2.5.4"
 
 __all__ = ["Error", "model", "read"]
 
@@ -65,7 +65,8 @@ def read(
     from prov.model import ProvDocument
     from prov.serializers import Registry
 
-    Registry.load_serializers()
+    if Registry.serializers is None:
+        Registry.load_serializers()
     assert Registry.serializers is not None  # populated by load_serializers()
     serializers = Registry.serializers.keys()
 

{prov-2.5.2 → prov-2.5.4}/src/prov/identifier.py

@@ -1,10 +1,19 @@
 from __future__ import annotations  # needed for | type annotations in Python < 3.10
 
-from typing import Any
+from typing import Any, ClassVar
 
 __author__ = "Trung Dong Huynh"
 __email__ = "trungdong@donggiang.com"
 
+# PROV-N metacharacters that must be backslash-escaped in the local part of a
+# qualified name (grammar production [55] PN_CHARS_ESC, #223). PN_CHARS_ESC
+# also lists '-' and '.' as escapable, but both are otherwise legal unescaped
+# there ('-' is a PN_CHARS character; '.' is legal unescaped except as the
+# final character of PN_LOCAL) so they are left untouched here -- only the
+# nine characters that are illegal unescaped anywhere in a local part are
+# escaped.
+_PROVN_LOCAL_ESCAPE = str.maketrans({c: f"\\{c}" for c in "='(),:;[]"})
+
 
 class Identifier:
     """Base class for all identifiers and also represents xsd:anyURI."""
@@ -91,14 +100,32 @@ class QualifiedName(Identifier):
     def __hash__(self) -> int:
         return hash(self.uri)
 
+    def provn_bare_representation(self) -> str:
+        """Return the ``prefix:local`` PROV-N form used at IDENTIFIER positions.
+
+        The local part's PROV-N metacharacters (``= ' ( ) , : ; [ ]``) are
+        backslash-escaped per grammar production [55] ``PN_CHARS_ESC`` (#223);
+        the prefix is never escaped, as it cannot contain these characters.
+        """
+        escaped_localpart = self._localpart.translate(_PROVN_LOCAL_ESCAPE)
+        return (
+            ":".join([self._namespace.prefix, escaped_localpart])
+            if self._namespace.prefix
+            else escaped_localpart
+        )
+
     def provn_representation(self) -> str:
         """Return the PROV-N representation of this qualified name as a quoted string."""
-        return f"'{self._str}'"
+        return f"'{self.provn_bare_representation()}'"
 
 
 class Namespace:
     """PROV Namespace."""
 
+    # _cache holds at most this many names. A lookup that misses when
+    # _cache is full clears it, then stores the new name.
+    _CACHE_LIMIT: ClassVar[int] = 10_000
+
     def __init__(self, prefix: str, uri: str):
         """Create a namespace with the given prefix and URI.
 
@@ -183,9 +210,11 @@ class Namespace:
         return f"<{self.__class__.__name__}: {self._prefix} {{{self._uri}}}>"
 
     def __getitem__(self, localpart: str) -> QualifiedName:
-        if localpart in self._cache:
-            return self._cache[localpart]
-        else:
+        cache = self._cache
+        qname = cache.get(localpart)
+        if qname is None:
             qname = QualifiedName(self, localpart)
-            self._cache[localpart] = qname
-            return qname
+            if len(cache) >= self._CACHE_LIMIT:
+                cache.clear()
+            cache[localpart] = qname
+        return qname

{prov-2.5.2 → prov-2.5.4}/src/prov/model/__init__.py

@@ -49,6 +49,7 @@ from prov.model.records import (
     AgentRef as AgentRef,
     DatetimeOrStr as DatetimeOrStr,
     EntityRef as EntityRef,
+    GenerationRef as GenerationRef,
     GenrationRef as GenrationRef,
     Literal as Literal,
     NameValuePair as NameValuePair,

{prov-2.5.2 → prov-2.5.4}/src/prov/model/bundle.py

@@ -75,7 +75,7 @@ from prov.model.records import (
     AgentRef,
     DatetimeOrStr,
     EntityRef,
-    GenrationRef,
+    GenerationRef,
     NameValuePair,
     NSCollection,
     OptionalID,
@@ -318,7 +318,11 @@ class ProvBundle:
 
         #  if this is the document, start the document;
         # otherwise, start the bundle
-        lines = ["document"] if self.is_document() else [f"bundle {self._identifier}"]
+        # #223: escape PN_CHARS_ESC metacharacters in the local part
+        bundle_id = (
+            self._identifier.provn_bare_representation() if self._identifier else ""
+        )
+        lines = ["document"] if self.is_document() else [f"bundle {bundle_id}"]
 
         default_namespace = self._namespaces.get_default_namespace()
         if default_namespace:
@@ -969,7 +973,7 @@ class ProvBundle:
         generatedEntity: EntityRef,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         identifier: OptionalID = None,
         other_attributes: RecordAttributesArg | None = None,
@@ -1012,7 +1016,7 @@ class ProvBundle:
         generatedEntity: EntityRef,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         identifier: OptionalID = None,
         other_attributes: RecordAttributesArg | None = None,
@@ -1057,7 +1061,7 @@ class ProvBundle:
         generatedEntity: EntityRef,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         identifier: OptionalID = None,
         other_attributes: RecordAttributesArg | None = None,
@@ -1102,7 +1106,7 @@ class ProvBundle:
         generatedEntity: EntityRef,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         identifier: OptionalID = None,
         other_attributes: RecordAttributesArg | None = None,

{prov-2.5.2 → prov-2.5.4}/src/prov/model/namespaces.py

@@ -3,6 +3,7 @@
 from __future__ import annotations  # needed for | type annotations in Python < 3.10
 
 from collections.abc import Iterable
+from typing import Any
 
 from prov.constants import PROV, XSD, XSI
 from prov.identifier import Identifier, Namespace, QualifiedName
@@ -12,7 +13,12 @@ DEFAULT_NAMESPACES = {"prov": PROV, "xsd": XSD, "xsi": XSI}
 
 
 class NamespaceManager(dict[str, Namespace]):
-    """Manages namespaces for PROV documents and bundles."""
+    """Manages namespaces for PROV documents and bundles.
+
+    Register namespaces through :meth:`add_namespace` and
+    :meth:`set_default_namespace`. Any of the dict's own write methods
+    bypasses the manager's index and is unsupported.
+    """
 
     parent = None  # type: NamespaceManager | None
     """Parent :class:`NamespaceManager` this manager is a child of, if any."""
@@ -36,6 +42,7 @@ class NamespaceManager(dict[str, Namespace]):
         dict.__init__(self)
         self._default_namespaces = DEFAULT_NAMESPACES
         self.update(self._default_namespaces)
+        self._reindex()
         self._namespaces = {}  # type: dict[str, Namespace]
 
         if default is not None:
@@ -83,6 +90,29 @@ class NamespaceManager(dict[str, Namespace]):
         """
         self._default = Namespace("", uri)
         self[""] = self._default
+        self._reindex()
+
+    def _index(self, namespace: Namespace) -> None:
+        by_uri = self._uris_by_length.setdefault(len(namespace.uri), {})
+        by_uri.setdefault(namespace.uri, []).append(namespace)
+
+    def _reindex(self) -> None:
+        """Rebuild the length index from the dict's namespaces.
+
+        The index holds every namespace the dict holds, by URI length and then
+        URI. The manager's writers keep it in step with the dict.
+        """
+        self._uris_by_length: dict[int, dict[str, list[Namespace]]] = {}
+        for namespace in self.values():
+            self._index(namespace)
+
+    def __setstate__(self, state: dict[str, Any]) -> None:
+        # The dict's items are already set when pickle calls this, so a state
+        # from a release without the index gets one built from them. A copy
+        # carries its own index and has no items yet.
+        self.__dict__.update(state)
+        if "_uris_by_length" not in state:
+            self._reindex()
 
     def get_default_namespace(self) -> Namespace | None:
         """Return the default namespace, or ``None`` if none is set."""
@@ -102,7 +132,8 @@ class NamespaceManager(dict[str, Namespace]):
             The registered namespace, which may be the existing or renamed one
             rather than the argument.
         """
-        if namespace in self.values():
+        # every namespace is stored under its own prefix
+        if self.get(namespace.prefix) == namespace:
             #  no need to do anything
             return namespace
         if namespace in self._rename_map:
@@ -134,6 +165,7 @@ class NamespaceManager(dict[str, Namespace]):
         # Only now add the namespace to the registry
         self._namespaces[prefix] = namespace
         self[prefix] = namespace
+        self._index(namespace)
         self._uri_map[uri] = namespace
 
         return namespace
@@ -232,10 +264,10 @@ class NamespaceManager(dict[str, Namespace]):
             else:
                 #  assuming it is a URI (with the first part as its scheme)
                 #  check if the URI can be compacted by any of the registered namespaces
-                for namespace in self.values():
-                    if str_value.startswith(namespace.uri):
-                        #  create a QName with the namespace
-                        return namespace[str_value.replace(namespace.uri, "")]
+                compacting = self._compacting_namespace(str_value)
+                if compacting is not None:
+                    #  create a QName with the namespace
+                    return compacting[str_value.replace(compacting.uri, "")]
         elif self._default and isinstance(qname, str):
             # no colon in the identifier and a default namespace is defined,
             # create and return a qualified name in the default namespace
@@ -249,6 +281,23 @@ class NamespaceManager(dict[str, Namespace]):
         # Default to FAIL
         return None
 
+    def _compacting_namespace(self, iri: str) -> Namespace | None:
+        """The first namespace in the dict's order whose URI prefixes ``iri``.
+
+        One candidate, the usual case, comes from the length index. Several
+        candidates fall back to a scan in the dict's order.
+        """
+        matches = [
+            namespace
+            for length, by_uri in self._uris_by_length.items()
+            for namespace in by_uri.get(iri[:length], ())
+        ]
+        candidates: Iterable[Namespace] = matches if len(matches) < 2 else self.values()
+        for namespace in candidates:
+            if iri.startswith(namespace.uri):
+                return namespace
+        return None
+
     def get_anonymous_identifier(self, local_prefix: str = "id") -> Identifier:
         """Return a fresh anonymous (blank-node) identifier.
 

{prov-2.5.2 → prov-2.5.4}/src/prov/model/records.py

@@ -88,7 +88,11 @@ OptionalID = QualifiedNameCandidate | None  # type: typing.TypeAlias
 EntityRef = Union["ProvEntity", QualifiedNameCandidate]  # type: typing.TypeAlias
 ActivityRef = Union["ProvActivity", QualifiedNameCandidate]  # type: typing.TypeAlias
 AgentRef = Union["ProvAgent", "ProvEntity", "ProvActivity", QualifiedNameCandidate]  # type: typing.TypeAlias
-GenrationRef = Union["ProvGeneration", QualifiedNameCandidate]  # type: typing.TypeAlias
+GenerationRef = Union["ProvGeneration", QualifiedNameCandidate]  # type: typing.TypeAlias
+#: Misspelling of :data:`GenerationRef`, kept as an alias because it has been
+#: the public spelling since 2.4.0 and the 2.x line does not remove documented
+#: names. `master` renames it outright for 3.0.
+GenrationRef = GenerationRef  # type: typing.TypeAlias
 UsageRef = Union["ProvUsage", QualifiedNameCandidate]  # type: typing.TypeAlias
 RecordAttributesArg = (
     dict[QualifiedNameCandidate, Any] | Iterable[tuple[QualifiedNameCandidate, Any]]
@@ -194,6 +198,9 @@ def first(a_set: set[Any]) -> Any | None:
 def _ensure_multiline_string_triple_quoted(value: str) -> str:
     # converting the value to a string
     s = str(value)
+    # Escape backslashes first, so quote-escaping below doesn't re-escape
+    # the backslashes it just introduced.
+    s = s.replace("\\", "\\\\")
     # Escaping any double quote
     s = s.replace('"', '\\"')
     if "\n" in s:
@@ -667,7 +674,8 @@ class ProvRecord:
         # Generating identifier
         relation_id = ""  # default blank
         if self._identifier:
-            identifier = str(self._identifier)  # TODO: QName export
+            # #223: escape PN_CHARS_ESC metacharacters in the local part
+            identifier = self._identifier.provn_bare_representation()
             if self.is_element():
                 items.append(identifier)
             else:
@@ -680,12 +688,13 @@ class ProvRecord:
             if values:
                 # Formal attributes always have single values
                 value = first(values)
-                # TODO: QName export
-                items.append(
-                    value.isoformat()
-                    if isinstance(value, datetime.datetime)
-                    else str(value)
-                )
+                if isinstance(value, datetime.datetime):
+                    items.append(value.isoformat())
+                elif isinstance(value, QualifiedName):
+                    # #223: escape PN_CHARS_ESC metacharacters in the local part
+                    items.append(value.provn_bare_representation())
+                else:
+                    items.append(str(value))
             else:
                 items.append("-")
 
@@ -699,8 +708,9 @@ class ProvRecord:
                         provn_represenation = value.provn_representation()
                     except AttributeError:
                         provn_represenation = encoding_provn_value(value)
-                    # TODO: QName export
-                    extra.append(f"{attr!s}={provn_represenation}")
+                    # #223: escape PN_CHARS_ESC metacharacters in the local part
+                    attr_name = attr.provn_bare_representation()
+                    extra.append(f"{attr_name}={provn_represenation}")
 
         if extra:
             # .format(), not an f-string: the nested string literals reuse the
@@ -817,7 +827,7 @@ class ProvEntity(ProvElement):
         self,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         attributes: RecordAttributesArg | None = None,
     ) -> ProvEntity:
@@ -900,7 +910,7 @@ class ProvEntity(ProvElement):
         self,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         attributes: RecordAttributesArg | None = None,
     ) -> ProvEntity:
@@ -929,7 +939,7 @@ class ProvEntity(ProvElement):
         self,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         attributes: RecordAttributesArg | None = None,
     ) -> ProvEntity:
@@ -958,7 +968,7 @@ class ProvEntity(ProvElement):
         self,
         usedEntity: EntityRef,
         activity: ActivityRef | None = None,
-        generation: GenrationRef | None = None,
+        generation: GenerationRef | None = None,
         usage: UsageRef | None = None,
         attributes: RecordAttributesArg | None = None,
     ) -> ProvEntity:

{prov-2.5.2 → prov-2.5.4}/src/prov/serializers/__init__.py

@@ -101,13 +101,12 @@ class Registry:
         (see :func:`get`) rather than a bare ``ModuleNotFoundError``.
 
         The insertion order is kept as ``json, rdf, provn, xml`` (the
-        historic order when all extras are present) because
-        :func:`prov.read`'s format auto-detection iterates
-        ``Registry.serializers`` in order and several tests
-        (``test_read_auto_detects_rdf``,
-        ``test_read_auto_detect_of_xml_hits_uncaught_rdf_syntax_error``,
-        ``test_read_on_unparseable_content_raises_bad_syntax``) pin the
-        exact candidate tried second.
+        historic order when all extras are present) because :func:`prov.read`'s
+        format auto-detection iterates ``Registry.serializers`` in order and
+        ``test_read_auto_detect_with_broken_tell_degrades_to_no_rewind``
+        pins JSON as the first candidate tried — a non-seekable stream can
+        only attempt the first format, so JSON must remain first for JSON
+        content to be auto-detected on broken streams.
         """
         from prov.serializers.provjson import ProvJSONSerializer
         from prov.serializers.provn import ProvNSerializer