[Findit] Add skeleton code for delta test script.

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

+# 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(
+      '--client',
+      '-c',
+      default='fracas',
+      help=('Possible values are: fracas, cracas, clusterfuzz. Right now, only '
+            'fracas is supported.'))
+
+  argparser.add_argument(
+      '--since',
+      '--after',

[stgao, 2016/10/14 01:40:40]

same as the other CL, why two different names?

[Sharu Jiang, 2016/10/15 01:24:47]

As answered in the other cl, this is similar to the args of 'git log', have two
options for people to remember :)

[stgao, 2016/10/20 01:40:20]

The other was updated. Why not here?
Two names for the same argument is uncommon.

[Sharu Jiang, 2016/10/20 22:39:06]

Done.

+      default=_A_YEAR_AGO,
+      help=('Query data since this date (including this date). '
+            'Should be in YYYY-MM-DD format. E.g. 2015-09-31. '
+            'Defaults to a year ago.'))
+
+  argparser.add_argument(
+      '--until',
+      '--before',
+      default=_TODAY,
+      help=('Query data until this date (not including this date). '
+            'Should be in YYYY-MM-DD format. E.g. 2015-09-31. '
+            'Defaults to today.'))
+
+  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.

[stgao, 2016/10/14 01:40:40]

Just Gqlquery?

[Sharu Jiang, 2016/10/15 01:24:47]

Oops, needs to be updated.

+  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])
+
+  delta_path = os.path.join(DELTA_RESULTS_DIRECTORY,
+                            '%s_%s_%s..%s.delta.csv' % (git_hash1[:7],
+                                                        git_hash2[:7],
+                                                        args.since, args.until))
+
+  # Check if delta results already existed.
+  if os.path.exists(delta_path):
+    logging.info('Delta results already existed in\n%s', delta_path)

[stgao, 2016/10/14 01:40:40]

For this case, what's the expected action by the user?

[Sharu Jiang, 2016/10/15 01:24:47]

In this case, user can check out the csv result file.

[stgao, 2016/10/20 01:40:20]

Why not add that the the message? It is not clear at least to me. And what do
you mean by "check out the csv result file"?

[Sharu Jiang, 2016/10/20 22:39:06]

I meant open the csv file to see the results. 

Modify the code so, even we already have csv result, we still print the delta to
the screen.

+    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, args.client,
+      start_date=args.since, end_date=args.until,
+      batch_size=args.batch, verbose=args.verbose,
+      app_id=os.getenv('APP_ID'))
+
+  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()