Improve performance of name normalization (#533)

Author: jaraco

exercises.py

@@ -45,3 +45,13 @@ def entrypoint_regexp_perf():
     input = '0' + ' ' * 2**10 + '0'  # end warmup
 
     re.match(importlib_metadata.EntryPoint.pattern, input)
+
+
+def normalize_perf():
+    # python/cpython#143658
+    import importlib_metadata  # end warmup
+
+    # operation completes in < 1ms, so repeat it to get visibility
+    # https://github.com/jaraco/pytest-perf/issues/12
+    for _ in range(1000):
+        importlib_metadata.Prepared.normalize('sample')

importlib_metadata/__init__.py

@@ -945,8 +945,15 @@ def __init__(self, name: str | None):
     def normalize(name):
         """
         PEP 503 normalization plus dashes as underscores.
+
+        Specifically avoids ``re.sub`` as prescribed for performance
+        benefits (see python/cpython#143658).
         """
-        return re.sub(r"[-_.]+", "-", name).lower().replace('-', '_')
+        value = name.lower().replace("-", "_").replace(".", "_")
+        # Condense repeats
+        while "__" in value:
+            value = value.replace("__", "_")
+        return value
 
     @staticmethod
     def legacy_normalize(name):

tests/test_api.py

@@ -6,6 +6,7 @@
 from importlib_metadata import (
     Distribution,
     PackageNotFoundError,
+    Prepared,
     distribution,
     entry_points,
     files,
@@ -317,3 +318,34 @@ class InvalidateCache(unittest.TestCase):
     def test_invalidate_cache(self):
         # No externally observable behavior, but ensures test coverage...
         importlib.invalidate_caches()
+
+
+class PreparedTests(unittest.TestCase):
+    @fixtures.parameterize(
+        # Simple
+        dict(input='sample', expected='sample'),
+        # Mixed case
+        dict(input='Sample', expected='sample'),
+        dict(input='SAMPLE', expected='sample'),
+        dict(input='SaMpLe', expected='sample'),
+        # Separator conversions
+        dict(input='sample-pkg', expected='sample_pkg'),
+        dict(input='sample.pkg', expected='sample_pkg'),
+        dict(input='sample_pkg', expected='sample_pkg'),
+        # Multiple separators
+        dict(input='sample---pkg', expected='sample_pkg'),
+        dict(input='sample___pkg', expected='sample_pkg'),
+        dict(input='sample...pkg', expected='sample_pkg'),
+        # Mixed separators
+        dict(input='sample-._pkg', expected='sample_pkg'),
+        dict(input='sample_.-pkg', expected='sample_pkg'),
+        # Complex
+        dict(input='Sample__Pkg-name.foo', expected='sample_pkg_name_foo'),
+        dict(input='Sample__Pkg.name__foo', expected='sample_pkg_name_foo'),
+        # Uppercase with separators
+        dict(input='SAMPLE-PKG', expected='sample_pkg'),
+        dict(input='Sample.Pkg', expected='sample_pkg'),
+        dict(input='SAMPLE_PKG', expected='sample_pkg'),
+    )
+    def test_normalize(self, input, expected):
+        self.assertEqual(Prepared.normalize(input), expected)