Speed up filename filtering.

Before there was a `getcwd` syscall for every filename which was filtered.

Instead this is now cached per-run.

- When all files are identified by filename only: ~45% improvement
- When no files are identified by filename only: ~55% improvement

This makes little difference to overall execution, the bigger win is
eliminating the `memoize_by_cwd` hack.  Just removing the memoization would
have *increased* the runtime by 300-500%.

pre_commit/meta_hooks/check_hooks_apply.py

@@ -3,24 +3,19 @@ import argparse
 import pre_commit.constants as C
 from pre_commit import git
 from pre_commit.clientlib import load_config
-from pre_commit.commands.run import _filter_by_include_exclude
-from pre_commit.commands.run import _filter_by_types
+from pre_commit.commands.run import Classifier
 from pre_commit.repository import all_hooks
 from pre_commit.store import Store
 
 
 def check_all_hooks_match_files(config_file):
-    files = git.get_all_files()
+    classifier = Classifier(git.get_all_files())
     retv = 0
 
     for hook in all_hooks(load_config(config_file), Store()):
         if hook.always_run or hook.language == 'fail':
             continue
-        include, exclude = hook.files, hook.exclude
-        filtered = _filter_by_include_exclude(files, include, exclude)
-        types, exclude_types = hook.types, hook.exclude_types
-        filtered = _filter_by_types(filtered, types, exclude_types)
-        if not filtered:
+        elif not classifier.filenames_for_hook(hook):
             print('{} does not apply to this repository'.format(hook.id))
             retv = 1
 

pre_commit/meta_hooks/check_useless_excludes.py

@@ -9,7 +9,7 @@ import pre_commit.constants as C
 from pre_commit import git
 from pre_commit.clientlib import load_config
 from pre_commit.clientlib import MANIFEST_HOOK_DICT
-from pre_commit.commands.run import _filter_by_types
+from pre_commit.commands.run import Classifier
 
 
 def exclude_matches_any(filenames, include, exclude):
@@ -24,11 +24,11 @@ def exclude_matches_any(filenames, include, exclude):
 
 def check_useless_excludes(config_file):
     config = load_config(config_file)
-    files = git.get_all_files()
+    classifier = Classifier(git.get_all_files())
     retv = 0
 
     exclude = config['exclude']
-    if not exclude_matches_any(files, '', exclude):
+    if not exclude_matches_any(classifier.filenames, '', exclude):
         print(
             'The global exclude pattern {!r} does not match any files'
             .format(exclude),
@@ -40,10 +40,11 @@ def check_useless_excludes(config_file):
             # Not actually a manifest dict, but this more accurately reflects
             # the defaults applied during runtime
             hook = apply_defaults(hook, MANIFEST_HOOK_DICT)
+            names = classifier.filenames
             types, exclude_types = hook['types'], hook['exclude_types']
-            filtered_by_types = _filter_by_types(files, types, exclude_types)
+            names = classifier.by_types(names, types, exclude_types)
             include, exclude = hook['files'], hook['exclude']
-            if not exclude_matches_any(filtered_by_types, include, exclude):
+            if not exclude_matches_any(names, include, exclude):
                 print(
                     'The exclude pattern {!r} for {} does not match any files'
                     .format(exclude, hook['id']),