Add --csv argument to find_unmapped_histograms.py

tools/metrics/histograms/find_unmapped_histograms.py

 # Copyright 2013 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.
 
 """Scans the Chromium source for histograms that are absent from histograms.xml.
 
 This is a heuristic scan, so a clean run of this script does not guarantee that
 all histograms in the Chromium source are properly mapped.  Notably, field
 trials are entirely ignored by this script.
 
@@ skipping 271 matching lines @@
 
   Args:
     name: The string to hash (a histogram name).
 
   Returns:
     Histogram hash as a string representing a hex number (with leading 0x).
   """
   return '0x' + hashlib.md5(name).hexdigest()[:16]
 
 
+def output_csv(unmapped_histograms, location_map):
+  for histogram in sorted(unmapped_histograms):
+    (filename, line_number) = location_map[histogram].split(':', 1)

[Ilya Sherman, 2016/08/17 22:09:24]

Hmm, is it really necessary to include the max-split arg of 1?  Presumably it
shouldn't be possible to have more than a single colon -- that's better handled
as an assertion, IMO.

+    print '%s,%s,%s,%s' % (filename, line_number, histogram,
+                           hashHistogramName(histogram))
+
+
+def output_log(unmapped_histograms, location_map, verbose):
+  if len(unmapped_histograms):
+    logging.info('')
+    logging.info('')
+    logging.info('Histograms in Chromium but not in XML files:')
+    logging.info('-------------------------------------------------')
+    for histogram in sorted(unmapped_histograms):
+      if verbose:
+        logging.info('%s: %s - %s', location_map[histogram], histogram,
+                     hashHistogramName(histogram))
+      else:
+        logging.info('  %s - %s', histogram, hashHistogramName(histogram))
+  else:
+    logging.info('Success!  No unmapped histograms found.')
+
+
 def main():
   # Find default paths.
   default_root = path_util.GetInputFile('/')
   default_histograms_path = path_util.GetInputFile(
       'tools/metrics/histograms/histograms.xml')
   default_extra_histograms_path = path_util.GetInputFile(
       'tools/histograms/histograms.xml')
 
   # Parse command line options
   parser = optparse.OptionParser()
   parser.add_option(
     '--root-directory', dest='root_directory', default=default_root,
     help='scan within DIRECTORY for histograms [optional, defaults to "%s"]' %
         default_root,
     metavar='DIRECTORY')
   parser.add_option(
     '--histograms-file', dest='histograms_file_location',
     default=default_histograms_path,
     help='read histogram definitions from FILE (relative to --root-directory) '
          '[optional, defaults to "%s"]' % default_histograms_path,
     metavar='FILE')
   parser.add_option(
     '--exrta_histograms-file', dest='extra_histograms_file_location',
     default=default_extra_histograms_path,
     help='read additional histogram definitions from FILE (relative to '
          '--root-directory) [optional, defaults to "%s"]' %
          default_extra_histograms_path,
     metavar='FILE')
   parser.add_option(
+      '--csv', action='store_true', dest='csv', default=False,

[Ilya Sherman, 2016/08/17 22:09:24]

nit: Let's name this variable 'output_as_csv' rather than just 'csv' (just for
the dest variable, not for the flag).

+      help=(
+          'output as csv for ease of parsing ' +
+          '[optional, defaults to %default]'))
+  parser.add_option(
       '--verbose', action='store_true', dest='verbose', default=False,
       help=(
           'print file position information with histograms ' +
           '[optional, defaults to %default]'))
 
   (options, args) = parser.parse_args()
   if args:
     parser.print_help()
     sys.exit(1)
 
   logging.basicConfig(format='%(levelname)s: %(message)s', level=logging.INFO)
 
   try:
     os.chdir(options.root_directory)
   except EnvironmentError as e:
     logging.error("Could not change to root directory: %s", e)
     sys.exit(1)
   chromium_histograms, location_map = readChromiumHistograms()
   xml_histograms = readXmlHistograms(options.histograms_file_location)
   unmapped_histograms = chromium_histograms - xml_histograms
 
   if os.path.isfile(options.extra_histograms_file_location):
     xml_histograms2 = readXmlHistograms(options.extra_histograms_file_location)
     unmapped_histograms -= xml_histograms2
   else:
     logging.warning('No such file: %s', options.extra_histograms_file_location)
 
-  if len(unmapped_histograms):
-    logging.info('')
-    logging.info('')
-    logging.info('Histograms in Chromium but not in XML files:')
-    logging.info('-------------------------------------------------')
-    for histogram in sorted(unmapped_histograms):
-      if options.verbose:
-        logging.info('%s: %s - %s', location_map[histogram], histogram,
-                     hashHistogramName(histogram))
-      else:
-        logging.info('  %s - %s', histogram, hashHistogramName(histogram))
+  if options.csv:
+    output_csv(unmapped_histograms, location_map)
   else:
-    logging.info('Success!  No unmapped histograms found.')
+    output_log(unmapped_histograms, location_map, options.verbose)
 
 
 if __name__ == '__main__':
   main()