https://github.com/python/cpython/commit/ee3657209b1464d66c89640e4db0dccdeec5b45c
commit: ee3657209b1464d66c89640e4db0dccdeec5b45c
branch: main
author: Adam Turner <9087854+aa-tur...@users.noreply.github.com>
committer: AA-Turner <9087854+aa-tur...@users.noreply.github.com>
date: 2025-04-08T10:05:48Z
summary:
gh-118761: Optimise import time for ``string`` (#132037)
Co-authored-by: Serhiy Storchaka <storch...@gmail.com>
files:
A Misc/NEWS.d/next/Library/2025-04-03-01-35-02.gh-issue-118761.VQcj70.rst
M Lib/string.py
diff --git a/Lib/string.py b/Lib/string.py
index c4f05c7223ce8a..eab5067c9b133e 100644
--- a/Lib/string.py
+++ b/Lib/string.py
@@ -49,11 +49,18 @@ def capwords(s, sep=None):
####################################################################
-import re as _re
-from collections import ChainMap as _ChainMap
-
_sentinel_dict = {}
+
+class _TemplatePattern:
+ # This descriptor is overwritten in ``Template._compile_pattern()``.
+ def __get__(self, instance, cls=None):
+ if cls is None:
+ return self
+ return cls._compile_pattern()
+_TemplatePattern = _TemplatePattern()
+
+
class Template:
"""A string class for supporting $-substitutions."""
@@ -64,14 +71,21 @@ class Template:
# See https://bugs.python.org/issue31672
idpattern = r'(?a:[_a-z][_a-z0-9]*)'
braceidpattern = None
- flags = _re.IGNORECASE
+ flags = None # default: re.IGNORECASE
+
+ pattern = _TemplatePattern # use a descriptor to compile the pattern
def __init_subclass__(cls):
super().__init_subclass__()
- if 'pattern' in cls.__dict__:
- pattern = cls.pattern
- else:
- delim = _re.escape(cls.delimiter)
+ cls._compile_pattern()
+
+ @classmethod
+ def _compile_pattern(cls):
+ import re # deferred import, for performance
+
+ pattern = cls.__dict__.get('pattern', _TemplatePattern)
+ if pattern is _TemplatePattern:
+ delim = re.escape(cls.delimiter)
id = cls.idpattern
bid = cls.braceidpattern or cls.idpattern
pattern = fr"""
@@ -82,7 +96,10 @@ def __init_subclass__(cls):
(?P<invalid>) # Other ill-formed delimiter exprs
)
"""
- cls.pattern = _re.compile(pattern, cls.flags | _re.VERBOSE)
+ if cls.flags is None:
+ cls.flags = re.IGNORECASE
+ pat = cls.pattern = re.compile(pattern, cls.flags | re.VERBOSE)
+ return pat
def __init__(self, template):
self.template = template
@@ -105,7 +122,8 @@ def substitute(self, mapping=_sentinel_dict, /, **kws):
if mapping is _sentinel_dict:
mapping = kws
elif kws:
- mapping = _ChainMap(kws, mapping)
+ from collections import ChainMap
+ mapping = ChainMap(kws, mapping)
# Helper function for .sub()
def convert(mo):
# Check the most common path first.
@@ -124,7 +142,8 @@ def safe_substitute(self, mapping=_sentinel_dict, /, **kws):
if mapping is _sentinel_dict:
mapping = kws
elif kws:
- mapping = _ChainMap(kws, mapping)
+ from collections import ChainMap
+ mapping = ChainMap(kws, mapping)
# Helper function for .sub()
def convert(mo):
named = mo.group('named') or mo.group('braced')
@@ -170,10 +189,6 @@ def get_identifiers(self):
self.pattern)
return ids
-# Initialize Template.pattern. __init_subclass__() is automatically called
-# only for subclasses, not for the Template class itself.
-Template.__init_subclass__()
-
########################################################################
# the Formatter class
diff --git
a/Misc/NEWS.d/next/Library/2025-04-03-01-35-02.gh-issue-118761.VQcj70.rst
b/Misc/NEWS.d/next/Library/2025-04-03-01-35-02.gh-issue-118761.VQcj70.rst
new file mode 100644
index 00000000000000..257ad7ece7d18a
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2025-04-03-01-35-02.gh-issue-118761.VQcj70.rst
@@ -0,0 +1,2 @@
+Improve import times by up to 27x for the :mod:`string` module.
+Patch by Adam Turner.
_______________________________________________
Python-checkins mailing list -- python-checkins@python.org
To unsubscribe send an email to python-checkins-le...@python.org
https://mail.python.org/mailman3/lists/python-checkins.python.org/
Member address: arch...@mail-archive.com
