Improve script to find user actions.

tools/metrics/actions/extract_actions.py

 #!/usr/bin/env python
 #
 # Copyright (c) 2012 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.
 
 """Extract UserMetrics "actions" strings from the Chrome source.
 
 This program generates the list of known actions we expect to see in the
 user behavior logs.  It walks the Chrome source, looking for calls to
 UserMetrics functions, extracting actions and warning on improper calls,
 as well as generating the lists of possible actions in situations where
 there are many possible actions.
 
 See also:
   base/metrics/user_metrics.h
-  http://wiki.corp.google.com/twiki/bin/view/Main/ChromeUserExperienceMetrics
 
 After extracting all actions, the content will go through a pretty print
 function to make sure it's well formatted. If the file content needs to be
 changed, a window will be prompted asking for user's consent. The old version
 will also be saved in a backup file.
 """
 
 __author__ = 'evanm (Evan Martin)'
 
 from HTMLParser import HTMLParser
 import logging
 import os
 import re
 import shutil
 import sys
 from xml.dom import minidom
 
 import print_style
 
 sys.path.insert(1, os.path.join(sys.path[0], '..', '..', 'python'))
 from google import path_utils
 
 # Import the metrics/common module for pretty print xml.
 sys.path.append(os.path.join(os.path.dirname(__file__), '..', 'common'))
 import diff_util
 import pretty_print_xml
 
+USER_METRICS_ACTION_RE = re.compile(r'UserMetricsAction\(\s*(.+?)\)')

[Ilya Sherman, 2015/01/09 22:25:43]

nit: I think it might be worth using the verbose regex format for this guy, and
document the individual pieces, since it's a little bit complex.

[Ilya Sherman, 2015/01/09 22:25:44]

It looks like you've changed the regexes quite a bit.  Why did you drop the
[^a-zA-Z] check, for example?  Also, are you sure that ".+" is correct?  I think
that would allow the match to go past any parens, e.g. to match 
"""UserMetricsAction("I'm an action!"), "I'm some other parameter.");"""

[Alexei Svitkine (slow), 2015/01/12 19:58:24]

I've re-added it now.
I added a test case for your example and it works correctly.

[Alexei Svitkine (slow), 2015/01/12 19:58:24]

Done.

+COMPUTED_ACTION_RE = re.compile(r'RecordComputedAction')
+QUOTED_STRING_RE = re.compile(r'\"(.+?)\"')
+
 # Files that are known to use content::RecordComputedAction(), which means
 # they require special handling code in this script.
 # To add a new file, add it to this list and add the appropriate logic to
 # generate the known actions to AddComputedActions() below.
 KNOWN_COMPUTED_USERS = (
   'back_forward_menu_model.cc',
   'options_page_view.cc',
   'render_view_host.cc',  # called using webkit identifiers
   'user_metrics.cc',  # method definition
   'new_tab_ui.cc',  # most visited clicks 1-9
@@ skipping 348 matching lines @@
   actions.add('ConnectivityDiagnostics.LaunchSource.WebStore')
   actions.add('ConnectivityDiagnostics.UA.LogsShown')
   actions.add('ConnectivityDiagnostics.UA.PassingTestsShown')
   actions.add('ConnectivityDiagnostics.UA.SettingsShown')
   actions.add('ConnectivityDiagnostics.UA.TestResultExpanded')
   actions.add('ConnectivityDiagnostics.UA.TestSuiteRun')
 
   # Actions sent by 'Ok Google' Hotwording.
   actions.add('Hotword.HotwordTrigger')
 
+def FindActionNames(contents, pos):
+  """Finds actions from the first UserMetricsAction() call in |contents|.
+
+  Arguments:
+    contents: string to search through
+    pos: position in |contents| to start the search from
+
+  Returns:
+    The list action names (strings) that was found.

[Ilya Sherman, 2015/01/09 22:25:43]

Please update this -- it looks like there are two values returned.

[Alexei Svitkine (slow), 2015/01/12 19:58:24]

Done.

+  """
+  match = USER_METRICS_ACTION_RE.search(contents, pos=pos)
+  if not match:
+    return None, None
+  return QUOTED_STRING_RE.findall(match.group(1)), match.end()
+
 def GrepForActions(path, actions):
   """Grep a source file for calls to UserMetrics functions.
 
   Arguments:
     path: path to the file
     actions: set of actions to add to
   """
   global number_of_files_total
   number_of_files_total = number_of_files_total + 1
-  # we look for the UserMetricsAction structure constructor
-  # this should be on one line
-  action_re = re.compile(r'[^a-zA-Z]UserMetricsAction\("([^"]*)')
-  malformed_action_re = re.compile(r'[^a-zA-Z]UserMetricsAction\([^"]')
-  computed_action_re = re.compile(r'RecordComputedAction')
+
+  contents = open(path).read()
+  pos = 0
+  while True:
+    action_names, pos = FindActionNames(contents, pos)
+    if not action_names:
+      break
+    actions.update(action_names)
+
   line_number = 0
   for line in open(path):
     line_number = line_number + 1
-    match = action_re.search(line)
-    if match:  # Plain call to RecordAction
-      actions.add(match.group(1))
-    elif malformed_action_re.search(line):
-      # Warn if this line is using RecordAction incorrectly.
-      print >>sys.stderr, ('WARNING: %s has malformed call to RecordAction'
-                           ' at %d' % (path, line_number))
-    elif computed_action_re.search(line):
+    if COMPUTED_ACTION_RE.search(line):
       # Warn if this file shouldn't be calling RecordComputedAction.
       if os.path.basename(path) not in KNOWN_COMPUTED_USERS:
         print >>sys.stderr, ('WARNING: %s has RecordComputedAction at %d' %
                              (path, line_number))
 
 class WebUIActionsParser(HTMLParser):
   """Parses an HTML file, looking for all tags with a 'metric' attribute.
   Adds user actions corresponding to any metrics found.
 
   Arguments:
@@ skipping 366 matching lines @@
 
   with open(actions_xml_path, 'wb') as f:
     f.write(pretty)
   print ('Updated %s. Don\'t forget to add it to your changelist' %
          actions_xml_path)
   return 0
 
 
 if '__main__' == __name__:
   sys.exit(main(sys.argv))