GH-101362: Optimise PurePath(PurePath(...)) (GH-101667)

The previous `_parse_args()` method pulled the `_parts` out of any supplied `PurePath` objects; these were subsequently joined in `_from_parts()` using `os.path.join()`. This is actually a slower form of joining than calling `fspath()` on the path object, because it doesn't take advantage of the fact that the contents of `_parts` is normalized!

This reduces the time taken to run `PurePath("foo", "bar")` by ~20%, and the time taken to run `PurePath(p, "cheese")`, where `p = PurePath("/foo", "bar", "baz")`, by ~40%.

Automerge-Triggered-By: GH:AlexWaygood
This commit is contained in:
Barney Gale 2023-03-05 23:50:21 +00:00 committed by GitHub
parent 3e60e0213e
commit 6716254e71
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 45 additions and 27 deletions

View file

@ -281,6 +281,14 @@ class PurePath(object):
path = cls._flavour.join(*parts)
sep = cls._flavour.sep
altsep = cls._flavour.altsep
if isinstance(path, str):
# Force-cast str subclasses to str (issue #21127)
path = str(path)
else:
raise TypeError(
"argument should be a str or an os.PathLike "
"object where __fspath__ returns a str, "
f"not {type(path).__name__!r}")
if altsep:
path = path.replace(altsep, sep)
drv, root, rel = cls._flavour.splitroot(path)
@ -291,32 +299,10 @@ class PurePath(object):
parsed = [sys.intern(x) for x in unfiltered_parsed if x and x != '.']
return drv, root, parsed
@classmethod
def _parse_args(cls, args):
# This is useful when you don't want to create an instance, just
# canonicalize some constructor arguments.
parts = []
for a in args:
if isinstance(a, PurePath):
parts += a._parts
else:
a = os.fspath(a)
if isinstance(a, str):
# Force-cast str subclasses to str (issue #21127)
parts.append(str(a))
else:
raise TypeError(
"argument should be a str object or an os.PathLike "
"object returning str, not %r"
% type(a))
return cls._parse_parts(parts)
@classmethod
def _from_parts(cls, args):
# We need to call _parse_args on the instance, so as to get the
# right flavour.
self = object.__new__(cls)
drv, root, parts = self._parse_args(args)
drv, root, parts = self._parse_parts(args)
self._drv = drv
self._root = root
self._parts = parts
@ -575,7 +561,7 @@ class PurePath(object):
anchored).
"""
drv1, root1, parts1 = self._drv, self._root, self._parts
drv2, root2, parts2 = self._parse_args(args)
drv2, root2, parts2 = self._parse_parts(args)
if root2:
if not drv2 and drv1:
return self._from_parsed_parts(drv1, root2, [drv1 + root2] + parts2[1:])
@ -662,7 +648,7 @@ class PurePath(object):
return True
# Can't subclass os.PathLike from PurePath and keep the constructor
# optimizations in PurePath._parse_args().
# optimizations in PurePath.__slots__.
os.PathLike.register(PurePath)