Replace custom wheel filename regex with packaging parse_wheel_filename

Author: notatallshaw

src/pip/_internal/index/package_finder.py

@@ -528,11 +528,7 @@ def _sort_key(self, candidate: InstallationCandidate) -> CandidateSortingKey:
                 )
             if self._prefer_binary:
                 binary_preference = 1
-            if wheel.build_tag is not None:
-                match = re.match(r"^(\d+)(.*)$", wheel.build_tag)
-                assert match is not None, "guaranteed by filename validation"
-                build_tag_groups = match.groups()
-                build_tag = (int(build_tag_groups[0]), build_tag_groups[1])
+            build_tag = wheel.build_tag
         else:  # sdist
             pri = -(support_num)
         has_allowed_hash = int(link.is_hash_allowed(self._hashes))

src/pip/_internal/metadata/importlib/_envs.py

@@ -8,10 +8,14 @@
 import zipimport
 from typing import Iterator, List, Optional, Sequence, Set, Tuple
 
-from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.utils import (
+    InvalidWheelFilename,
+    NormalizedName,
+    canonicalize_name,
+    parse_wheel_filename,
+)
 
 from pip._internal.metadata.base import BaseDistribution, BaseEnvironment
-from pip._internal.models.wheel import Wheel
 from pip._internal.utils.deprecation import deprecated
 from pip._internal.utils.filetypes import WHEEL_EXTENSION
 
@@ -26,7 +30,9 @@ def _looks_like_wheel(location: str) -> bool:
         return False
     if not os.path.isfile(location):
         return False
-    if not Wheel.wheel_file_re.match(os.path.basename(location)):
+    try:
+        parse_wheel_filename(os.path.basename(location))
+    except InvalidWheelFilename:
         return False
     return zipfile.is_zipfile(location)
 

src/pip/_internal/models/wheel.py

@@ -2,45 +2,29 @@
 name that have meaning.
 """
 
-import re
 from typing import Dict, Iterable, List
 
 from pip._vendor.packaging.tags import Tag
+from pip._vendor.packaging.utils import (
+    InvalidWheelFilename as _PackagingInvalidWheelFilename,
+)
+from pip._vendor.packaging.utils import parse_wheel_filename
 
 from pip._internal.exceptions import InvalidWheelFilename
 
 
 class Wheel:
     """A wheel file"""
 
-    wheel_file_re = re.compile(
-        r"""^(?P<namever>(?P<name>[^\s-]+?)-(?P<ver>[^\s-]*?))
-        ((-(?P<build>\d[^-]*?))?-(?P<pyver>[^\s-]+?)-(?P<abi>[^\s-]+?)-(?P<plat>[^\s-]+?)
-        \.whl|\.dist-info)$""",
-        re.VERBOSE,
-    )
-
     def __init__(self, filename: str) -> None:
-        """
-        :raises InvalidWheelFilename: when the filename is invalid for a wheel
-        """
-        wheel_info = self.wheel_file_re.match(filename)
-        if not wheel_info:
-            raise InvalidWheelFilename(f"{filename} is not a valid wheel filename.")
         self.filename = filename
-        self.name = wheel_info.group("name").replace("_", "-")
-        # we'll assume "_" means "-" due to wheel naming scheme
-        # (https://github.com/pypa/pip/issues/1150)
-        self.version = wheel_info.group("ver").replace("_", "-")
-        self.build_tag = wheel_info.group("build")
-        self.pyversions = wheel_info.group("pyver").split(".")
-        self.abis = wheel_info.group("abi").split(".")
-        self.plats = wheel_info.group("plat").split(".")
-
-        # All the tag combinations from this file
-        self.file_tags = {
-            Tag(x, y, z) for x in self.pyversions for y in self.abis for z in self.plats
-        }
+        try:
+            wheel_info = parse_wheel_filename(filename)
+        except _PackagingInvalidWheelFilename as e:
+            raise InvalidWheelFilename(e.args[0]) from None
+
+        self.name, _version, self.build_tag, self.file_tags = wheel_info
+        self.version = str(_version)
 
     def get_formatted_file_tags(self) -> List[str]:
         """Return the wheel's tags as a sorted list of strings."""

tests/unit/test_models_wheel.py

@@ -11,35 +11,43 @@ def test_std_wheel_pattern(self) -> None:
         w = Wheel("simple-1.1.1-py2-none-any.whl")
         assert w.name == "simple"
         assert w.version == "1.1.1"
-        assert w.pyversions == ["py2"]
-        assert w.abis == ["none"]
-        assert w.plats == ["any"]
+        assert w.build_tag == ()
+        assert w.file_tags == frozenset(
+            [Tag(interpreter="py2", abi="none", platform="any")]
+        )
 
     def test_wheel_pattern_multi_values(self) -> None:
         w = Wheel("simple-1.1-py2.py3-abi1.abi2-any.whl")
         assert w.name == "simple"
         assert w.version == "1.1"
-        assert w.pyversions == ["py2", "py3"]
-        assert w.abis == ["abi1", "abi2"]
-        assert w.plats == ["any"]
+        assert w.build_tag == ()
+        assert w.file_tags == frozenset(
+            [
+                Tag(interpreter="py2", abi="abi1", platform="any"),
+                Tag(interpreter="py2", abi="abi2", platform="any"),
+                Tag(interpreter="py3", abi="abi1", platform="any"),
+                Tag(interpreter="py3", abi="abi2", platform="any"),
+            ]
+        )
 
     def test_wheel_with_build_tag(self) -> None:
         # pip doesn't do anything with build tags, but theoretically, we might
         # see one, in this case the build tag = '4'
         w = Wheel("simple-1.1-4-py2-none-any.whl")
         assert w.name == "simple"
         assert w.version == "1.1"
-        assert w.pyversions == ["py2"]
-        assert w.abis == ["none"]
-        assert w.plats == ["any"]
+        assert w.build_tag == (4, "")
+        assert w.file_tags == frozenset(
+            [Tag(interpreter="py2", abi="none", platform="any")]
+        )
 
     def test_single_digit_version(self) -> None:
         w = Wheel("simple-1-py2-none-any.whl")
         assert w.version == "1"
 
     def test_non_pep440_version(self) -> None:
-        w = Wheel("simple-_invalid_-py2-none-any.whl")
-        assert w.version == "-invalid-"
+        with pytest.raises(InvalidWheelFilename):
+            Wheel("simple-_invalid_-py2-none-any.whl")
 
     def test_missing_version_raises(self) -> None:
         with pytest.raises(InvalidWheelFilename):
@@ -172,8 +180,7 @@ def test_support_index_min__none_supported(self) -> None:
 
     def test_version_underscore_conversion(self) -> None:
         """
-        Test that we convert '_' to '-' for versions parsed out of wheel
-        filenames
+        '_' is not PEP 440 compliant in the version part of a wheel filename
         """
-        w = Wheel("simple-0.1_1-py2-none-any.whl")
-        assert w.version == "0.1-1"
+        with pytest.raises(InvalidWheelFilename):
+            Wheel("simple-0.1_1-py2-none-any.whl")