[Findit] Add skeleton code for delta test script.

appengine/findit/util_scripts/crash_queries/delta_test/run-delta-test.py

Index: appengine/findit/util_scripts/crash_queries/delta_test/run-delta-test.py
diff --git a/appengine/findit/util_scripts/crash_queries/delta_test/run-delta-test.py b/appengine/findit/util_scripts/crash_queries/delta_test/run-delta-test.py
new file mode 100644
index 0000000000000000000000000000000000000000..55bc638d70d8ca80910b150f89d36fef58244a57
--- /dev/null
+++ b/appengine/findit/util_scripts/crash_queries/delta_test/run-delta-test.py
@@ -0,0 +1,131 @@
+# Copyright 2016 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.
+
+"""Runs delta test on 2 findit versions."""
+
+import argparse
+from datetime import date
+from datetime import timedelta
+import logging
+import os
+import sys
+
+from crash.type_enums import CrashClient
+from crash_queries.delta_test import delta_test
+from crash_queries.delta_test import delta_util
+
+_TODAY = date.today().strftime('%Y-%m-%d')
+_A_YEAR_AGO = (date.today() - timedelta(days=365)).strftime('%Y-%m-%d')
+
+# App Engine APIs will fail if batch size is more than 1000.
+_MAX_BATCH_SIZE = 1000
+_DEFAULT_BATCH_SIZE = _MAX_BATCH_SIZE
+
+DELTA_RESULTS_DIRECTORY = os.path.join(os.path.dirname(__file__),
+                                       'delta_results')
+CHROMIUM_REPO = 'https://chromium.googlesource.com/chromium/src'
+
+
+def RunDeltaTest():
+  """Runs delta testing between 2 different Findit versions."""
+  argparser = argparse.ArgumentParser(
+      description='Run delta test between 2 findit versions.')
+
+  argparser.add_argument(
+      '--revisions',
+      '-r',
+      nargs='+',
+      default=['HEAD^', 'HEAD'],
+      help=('2 findit revisions to be compared. It can take '
+            '1 or 2 revisions.\n'
+            '(1)-r rev1 rev2: compare rev1 and rev2\n'
+            '(2)-r rev: compare rev and current HEAD\n'
+            '(3)no revisions provided, default to compare '
+            'HEAD^ and HEAD'))
+
+  argparser.add_argument(
+      '--date',
+      '-d',
+      default='%s..%s' % (_A_YEAR_AGO, _TODAY),
+      help=('Should be in start_date..end_date format.'
+            'If it\' 2016-08-01.., default end_date to today, if it\'s '
+            '..2016-08-16, default start_date to a year ago.'
+            'Start date and end date to query data in [start_date, end_date)'))
+
+  argparser.add_argument(
+      '--batch',
+      '-b',
+      type=int,
+      default=_DEFAULT_BATCH_SIZE,
+      help=('The size of batch that can be processed at one time.\n'
+            'Note, the batch size cannot be greater than 1000, or app engine '
+            'APIs would fail.'))
+
+  argparser.add_argument(
+      '--verbose',
+      '-v',
+      action='store_true',
+      default=False,
+      help='Print findit results.')
+
+  args = argparser.parse_args()
+
+  # If in verbose mode, prints debug infos about Gqlquery.
+  if args.verbose:
+    logging.basicConfig(level=logging.DEBUG)
+  else:
+    logging.basicConfig(level=logging.INFO)
+
+  if len(args.revisions) > 2:
+    logging.error('Only support delta test between 2 versions.')
+    sys.exit(1)
+
+  if args.batch > _MAX_BATCH_SIZE:
+    logging.error('Batch size cannot be greater than %s, or app engine APIs '
+                  'would fail.', _MAX_BATCH_SIZE)
+    sys.exit(1)
+
+  # If only one revision provided, default the rev2 to HEAD.
+  if len(args.revisions) == 1:
+    args.revisions.append('HEAD')
+
+  git_hash1 = delta_util.ParseGitHash(args.revisions[0])
+  git_hash2 = delta_util.ParseGitHash(args.revisions[1])
+
+  if '..' not in args.date:
+    logging.error('Invalid format for date')
+    sys.exit(1)
+
+  start_date, end_date = args.date.split('..')
+  if not start_date:
+    start_date = _A_YEAR_AGO
+  if not end_date:
+    end_date = _TODAY
+
+  delta_path = os.path.join(DELTA_RESULTS_DIRECTORY,
+                            '%s_%s_%s..%s.delta.csv' % (git_hash1[:7],
+                                                        git_hash2[:7],
+                                                        start_date, end_date))
+
+  # Check if delta results already existed.
+  if os.path.exists(delta_path):
+    logging.info('Delta results already existed in\n%s', delta_path)
+    return
+
+
+  logging.info('Running delta test...')
+  # Get delta of results between git_hash1 and git_hash2.
+  deltas, crash_num = delta_test.DeltaEvaluator(
+      git_hash1, git_hash2, CrashClient.FRACAS,
+      start_date, end_date,
+      args.batch, verbose=args.verbose)
+
+  logging.info('\n========================= Summary =========================')
+  delta_util.PrintDelta(deltas, crash_num)
+  delta_util.WriteDeltaToCSV(deltas, crash_num,
+                             git_hash1, git_hash2, delta_path)
+
+
+if __name__ == '__main__':
+  RunDeltaTest()