filter-repo: add basic path filtering

Signed-off-by: Elijah Newren <newren@gmail.com>

git-filter-repo

@@ -12,6 +12,7 @@ in order to make it into a well-rounded filtering tool.
 from __future__ import print_function
 
 import argparse
+import fnmatch
 import os
 import re
 import subprocess
@@ -1055,6 +1056,13 @@ _CURRENT_STREAM_NUMBER = 0
 
 ######################################################################
 
+class AppendFilter(argparse.Action):
+  def __call__(self, parser, namespace, values, option_string=None):
+    filter_type = option_string[len('--path-'):] or 'match'
+    items = getattr(namespace, self.dest, []) or []
+    items.append((mod_type, match_type, values))
+    setattr(namespace, self.dest, items)
+
 def get_args():
   parser = argparse.ArgumentParser(description='Rewrite repository history')
   # FIXME: Need to special case all --* args that rev-list takes, or call
@@ -1062,6 +1070,32 @@ def get_args():
   parser.add_argument('--force', '-f', action='store_true',
                       help='''Rewrite history even if the current repo does not
                               look like a fresh clone.''')
+
+  path_group = parser.add_argument_group(title='Filtering based on paths')
+  path_group.add_argument('--invert-paths', action='store_false',
+                          dest='inclusive',
+                          help='''Invert the selection of files from the
+                                  specified --path-{match,glob,regex} options
+                                  below, i.e. only select files matching none
+                                  of those options.''')
+
+  path_group.add_argument('--path-match', '--path', metavar='DIR_OR_FILE',
+                          action=AppendFilter, dest='path_filter',
+                          help='''Exact paths (files or directories) to include
+                                  in filtered history.  Multiple --path
+                                  options can be specified to get a union of
+                                  paths.''')
+  path_group.add_argument('--path-glob', metavar='GLOB',
+                          action=AppendFilter, dest='path_filter',
+                          help='''Glob of paths to include in filtered
+                                  history.  Multiple --path-glob options can
+                                  be specified to get a union of paths.''')
+  path_group.add_argument('--path-regex', metavar='REGEX',
+                          action=AppendFilter, dest='path_filter',
+                          help='''Regex of paths to include in filtered
+                                  history.  Multiple --path-regex options can
+                                  be specified to get a union of paths''')
+
   parser.add_argument('revisions', nargs='*',
                       help='''Branches/tags/refs to rewrite.  Special rev-list
                               options, such as --branches, --tags, --all,
@@ -1073,6 +1107,9 @@ def get_args():
   args = parser.parse_args()
   if not args.revisions:
     args.revisions = ['--all']
+  if args.path_filter == None:
+    args.path_filter = []
+    args.inclusive = False
   return args
 
 def is_repository_bare():
@@ -1144,6 +1181,28 @@ def get_refs():
   output = subprocess.check_output('git show-ref'.split())
   return dict(reversed(x.split()) for x in output.splitlines())
 
+def tweak_commit(args, commit):
+  def include_file(path_filter, pathname):
+    for (filter_type, path_expression) in path_filter:
+      if filter_type == 'match':
+        n = len(path_expression)
+        if (pathname.startswith(path_expression) and
+            (path_expression[n-1] == '/' or
+             len(pathname) == n or
+             pathname[n] == '/')):
+          return True
+      if filter_type == 'glob' and fnmatch.fnmatch(pathname, path_expression):
+        return True
+      if filter_type == 'regex' and re.search(path_expression, pathname):
+        return True
+    return False
+
+  new_file_changes = []
+  for change in commit.file_changes:
+    if include_file(args.path_filter, change.filename) == args.inclusive:
+      new_file_changes.append(change)
+  commit.file_changes = new_file_changes
+
 def run_fast_filter():
   args = get_args()
   orig_refs = get_refs()
@@ -1156,7 +1215,9 @@ def run_fast_filter():
                          stdout=subprocess.PIPE)
   fip = subprocess.Popen('git fast-import --force --quiet'.split(),
                          stdin=subprocess.PIPE)
-  filter = FastExportFilter()
+  filter = FastExportFilter(
+             commit_callback = lambda c : tweak_commit(args, c),
+             )
   filter.run(fep.stdout, fip.stdin)
   fip.stdin.close()
   if fep.wait():