Added script to generate stats on a git checkout

infra/tools/antibody/git_commit_parser.py

 # Copyright 2015 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 import datetime
 import dateutil.parser
 import pytz
 import re
 import subprocess
 
 import infra.tools.antibody.cloudsql_connect as csql
 
 curr_time = datetime.datetime.now()
 
 
 def read_commit_info(git_checkout_path, commits_after_date,
                      git_log_format=('%H', '%b', '%ae',
                                      '%ci')):  # pragma: no cover
   """Read commit messages and other information
 
   Args:
     git_checkout_path(str): path to a local git checkout
     git_log_format(str): formatting directives passed to git log --format
+
   Return:
     log(str): output of git log
   """
   git_log_format = '%x1f'.join(git_log_format) + '%x1e'
   log = subprocess.check_output(['git', 'log',
       '--format=%s' % git_log_format, '--after=%s' % commits_after_date],
       cwd=git_checkout_path)
   return log
 
 
@@ skipping 24 matching lines @@
                or re.match(r'^BUG=https?://crbug.com/(\d+)', git_line)
                or re.match(r'^BUG=chromium:(\d+)', git_line)
                or re.match(r'^BUG=(\d+)', git_line))
   if bug_match:
     bug_url = bug_match.group(1)
   return bug_url
 
 
 def get_tbr(git_line):
   tbr = None
-  if git_line.startswith('TBR=') and len(git_line) > 4:
-    tbr = git_line[4:]
-    tbr = [x.strip() for x in tbr.split(',')]
+  if git_line.startswith('TBR='):
+    if len(git_line) > 4:
+      tbr = git_line[4:]
+      tbr = [x.strip().split('@')[0] for x in tbr.split(',')]
+    else:
+      tbr = ['NOBODY']
   return tbr
 
 
+# TODO(keelerh): figure out how to parse existence of a review url if
+# not prefaced by any indicator (because review urls for other commits
+# appear in the commit message frequently for reverts and references)
 def get_review_url(git_line):
   review_url = None
   if re.match(r'^Review:.+$', git_line):
     review_url = git_line[8:]
   elif re.match(r'^Review URL:.+$', git_line):
     review_url = git_line[12:]
   elif re.match(r'^Code review URL:.+$', git_line):
     review_url = git_line[17:]
   return review_url
 
@@ skipping 21 matching lines @@
 def get_features_for_commit_people(git_commit):
   """Retrieves the people associated with a git commit
 
   Arg:
     git_commit(dict): a commit message parsed into a dictionary
 
   Return:
     (tuple): relevant people and type extracted from the commit
   """
   git_hash = git_commit['id']
-  author = git_commit['author']
+  author = git_commit['author'].split('@')[0]
   people_rows = [(author, git_hash, curr_time, 'author')]
   TBR = None
   for line in git_commit['body'].split('\n'):
     TBR = get_tbr(line) or TBR
   if TBR is not None:
     for person in TBR:
       people_rows.append((person, git_hash, curr_time, 'tbr'))
   return people_rows
 
 
@@ skipping 64 matching lines @@
       count(*) as c FROM commit_people WHERE type='tbr'
       GROUP BY git_commit_hash) tbr_count
       ON commit_people.git_commit_hash = tbr_count.git_commit_hash
       INNER JOIN git_commit
       ON commit_people.git_commit_hash = git_commit.hash
       WHERE tbr_count.c <> 0
       AND git_commit.review_url IS NOT NULL
       AND commit_people.type='author'""")
   commits_with_review_urls = cc.fetchall()
   return [x[0] for x in commits_with_review_urls]