Improve performance on pre-commit -r pyflakes by 40%

pre_commit/git.py

@@ -1,14 +1,16 @@
 import os
 import pkg_resources
-
 from plumbum import local
 
+from pre_commit.util import memoize_by_cwd
 
-# TODO: optimization: memoize based on local.cwd.getpath()
+
+@memoize_by_cwd
 def get_root():
     return local['git']['rev-parse', '--show-toplevel']().strip()
 
 
+@memoize_by_cwd
 def get_pre_commit_path():
     return os.path.join(get_root(), '.git/hooks/pre-commit')
 

pre_commit/util.py

@@ -1,4 +1,8 @@
 
+import functools
+import os
+
+
 class cached_property(object):
     """Like @property, but caches the value."""
 
@@ -14,3 +18,19 @@ class cached_property(object):
         value = self._func(obj)
         obj.__dict__[self.__name__] = value
         return value
+
+
+def memoize_by_cwd(func):
+    """Memoize a function call based on os.getcwd()."""
+    cache = {}
+    @functools.wraps(func)
+    def wrapper(*args):
+        cwd = os.getcwd()
+        key = (cwd,) + args
+        try:
+            return cache[key]
+        except KeyError:
+            ret = cache[key] = func(*args)
+            return ret
+
+    return wrapper

tests/util_test.py

@@ -1,8 +1,10 @@
 
 import pytest
-import time
+import random
+from plumbum import local
 
 from pre_commit.util import cached_property
+from pre_commit.util import memoize_by_cwd
 
 
 @pytest.fixture
@@ -10,7 +12,7 @@ def class_with_cached_property():
     class Foo(object):
         @cached_property
         def foo(self):
-            return "Foo" + str(time.time())
+            return "Foo" + str(random.getrandbits(64))
 
     return Foo
 
@@ -27,3 +29,33 @@ def test_unbound_cached_property(class_with_cached_property):
     prop = class_with_cached_property.foo
     assert isinstance(prop, cached_property)
 
+
+@pytest.fixture
+def memoized_by_cwd():
+    @memoize_by_cwd
+    def func(arg):
+        return arg + str(random.getrandbits(64))
+
+    return func
+
+
+def test_memoized_by_cwd_returns_same_twice_in_a_row(memoized_by_cwd):
+    ret = memoized_by_cwd('baz')
+    ret2 = memoized_by_cwd('baz')
+    assert ret is ret2
+
+
+def test_memoized_by_cwd_returns_different_for_different_args(memoized_by_cwd):
+    ret = memoized_by_cwd('baz')
+    ret2 = memoized_by_cwd('bar')
+    assert ret.startswith('baz')
+    assert ret2.startswith('bar')
+    assert ret != ret2
+
+
+def test_memoized_by_cwd_changes_with_different_cwd(memoized_by_cwd):
+    ret = memoized_by_cwd('baz')
+    with local.cwd('.git'):
+        ret2 = memoized_by_cwd('baz')
+
+    assert ret != ret2