Add script to find unused Polymer elements

third_party/polymer/v1_0/find_unused_elements.py

+#!/usr/bin/python2
+
+# 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.
+
+# 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.
+
+"""Identifies Polymer elements that downloaded but not used by Chrome.
+
+Only finds "first-order" unused elements; re-run after removing unused elements
+to check if other elements have become unused.
+"""
+
+import os
+import pyparsing

[stevenjb, 2016/04/29 20:34:56]

Used?

[michaelpg, 2016/05/02 19:31:45]

Done.

+import re
+import subprocess
+import sys

[stevenjb, 2016/04/29 20:34:55]

Used?

[michaelpg, 2016/05/02 19:31:45]

Done.

+
+
+class UnusedElementsDetector:

[Dan Beam, 2016/05/03 19:22:08]

this should probably inherit from (object)

[michaelpg, 2016/05/03 21:29:19]

Done.

+  """Finds unused Polymer elements."""
+
+  # Unused elements to ignore because we plan to use them soon.
+  __whitelist = (

[Dan Beam, 2016/05/03 19:22:08]

nit: I think it's more common to use 4\s indent in tuple literals

[Dan Beam, 2016/05/03 19:22:08]

nit: __WHITELIST (Class contant
https://google.github.io/styleguide/pyguide.html#Naming)

[michaelpg, 2016/05/03 21:29:19]

Done.

[michaelpg, 2016/05/03 21:29:19]

the examples in PEP-8 use 4 spaces (as with other indents), the style guide says
to use 2 spaces instead of 4 spaces, so that sounds "incorrect"

+    # TODO(dschuyler): Use element or remove from whitelist.
+    'carbon-route',
+    # TODO(tsergeant): Use element or remove from whitelist.
+    'iron-scroll-threshold',
+    # Necessary for closure.
+    'polymer-externs',
+  )
+
+  @staticmethod
+  def __StripHtmlComments(filename):
+    """Returns the contents of an HTML file with <!-- --> comments stripped.
+
+    Not a real parser.
+
+    Args:
+      filename: The name of the file to read.
+
+    Returns:
+      A string consisting of the file contents with comments removed.
+    """
+    text = open(filename).read()
+    return re.sub('<!--.*?-->', '', text, flags=re.MULTILINE|re.DOTALL)

[Dan Beam, 2016/05/03 19:22:08]

nit: flags=re.MULTILINE | re.DOTALL)
                       ^ ^

[michaelpg, 2016/05/03 21:29:19]

Done. Looked weird to me inside a param=value expression with no spaces.

+
+  @staticmethod
+  def __StripJsComments(filename):
+    """Returns the minified contents of a JavaScript file with comments and
+    grit directives removed.
+
+    Args:
+      filename: The name of the file to read.
+
+    Returns:
+      A string consisting of the minified file contents with comments and grit
+      directives removed.
+    """
+    text = open(filename).read()
+    text = re.sub('<if .*?>', '', text)

[Dan Beam, 2016/05/03 19:22:07]

nit: case-insensitive maybe?

[michaelpg, 2016/05/03 21:29:19]

Done.

+    text = re.sub('</if>', '', text)
+
+    proc = subprocess.Popen(['uglifyjs', filename], stdout=subprocess.PIPE)
+    return proc.stdout.read()
+
+  @staticmethod
+  def __StripComments(filename):
+    """Returns the contents of a JavaScript or HTML file with comments removed.
+
+    Args:
+      filename: The name of the file to read.
+
+    Returns:
+      A string consisting of the file contents processed via
+      __StripHtmlComments or __StripJsComments.
+    """
+    if filename.endswith('.html'):
+      text = UnusedElementsDetector.__StripHtmlComments(filename)
+    elif filename.endswith('.js'):
+      text = UnusedElementsDetector.__StripJsComments(filename)
+    else:
+      assert False, 'Invalid filename: %s' % filename
+    return text
+
+  @staticmethod
+  def Run(): 
+    """Finds unused Polymer elements and prints a summary."""
+    proc = subprocess.Popen(
+      ['git', 'rev-parse', '--show-toplevel'],
+      stdout=subprocess.PIPE)
+    src_dir  = proc.stdout.read().strip()
+    polymer_dir = os.path.dirname(os.path.realpath(__file__))
+    components_dir = os.path.join(polymer_dir, 'components-chromium')
+
+    elements = []
+    for name in os.listdir(components_dir):
+      path = os.path.join(components_dir, name)
+      if os.path.isdir(path):
+        elements.append(name)
+
+    unused_elements = []
+    for element in elements:
+      if element in UnusedElementsDetector.__whitelist:
+        continue
+
+      relevant_src_dirs = (
+        os.path.join(src_dir, 'chrome'),
+        os.path.join(src_dir, 'ui'),
+        os.path.join(src_dir, 'components'),
+        components_dir
+      )
+      used = False
+      for path in relevant_src_dirs:
+        # Find an import or script referencing the tag's directory.
+        for (dirpath, _, filenames) in os.walk(path):
+          # Ignore the element's own files.
+          if dirpath.startswith(os.path.join(components_dir, element)):
+            continue
+
+          for filename in [f for f in filenames if f.endswith('.html') or
+            f.endswith('.js')]:
+            # Skip generated files that may include the element source.
+            if filename in ('crisper.js', 'vulcanized.html'):
+              continue
+
+            text = open(os.path.join(dirpath, filename)).read()
+            if not re.search('/%s/' % element, text):
+              continue
+
+            # Check the file again, ignoring comments (e.g. example imports and
+            # scripts).
+            if re.search(
+              '/%s' % element,

[stevenjb, 2016/04/29 20:34:56]

My python formatting fu is weak, but I suspect this should be on the previous
line with the other arguments aligned.

[michaelpg, 2016/05/02 19:31:45]

Done (both are valid, but your way looks better)

+              UnusedElementsDetector.__StripComments(
+                os.path.join(dirpath, filename))):
+              used = True
+              break
+          if used:
+            break
+        if used:
+          break

[stevenjb, 2016/04/29 20:34:55]

This would be more readable if the entire for loop were wrapped as a function
with 'return True' instead of 'used = True'.

[michaelpg, 2016/05/02 19:31:45]

Meant to do that, forgot, done now. Thanks!

+      if not used:
+        unused_elements.append(element)
+
+    if len(unused_elements):
+      print 'Found unused elements: %s\nRemove from bower.json and re-run ' \
+        'reproduce.sh, or add to whitelist in %s' % (
+          ', '.join(unused_elements), os.path.basename(__file__))
+
+
+UnusedElementsDetector.Run()

[Dan Beam, 2016/05/03 19:22:07]

if __name__ == '__main__':
  UnusedElementDetector().run()

[michaelpg, 2016/05/03 21:29:19]

Done.