Extensions Docs Server: HTML parser in IDS

chrome/common/extensions/docs/server2/intro_data_source.py

 # 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.
 
-import re
+from HTMLParser import HTMLParser
+
 from path_utils import FormatKey
 from third_party.handlebar import Handlebar
 
+class _IntroParser(HTMLParser):
+  """ An HTML parser which will parse table of contents and page title info out
+  of an intro.
+  """
+  def init(self):
+    """ This method is needed because HTMLParser is an old style class that does
+    not inherit from |object|, so the super constructor cannot be called in
+    |__init__|.
+    """

[not at google - send to devlin, 2012/07/23 12:47:36]

I think the pattern is like

def __init_(self):
  HTMLParser.__init__(self)
  self.toc = []
  ...

[cduvall, 2012/07/23 20:24:21]

Done.

+    self.toc = []
+    self.page_title = ''

[not at google - send to devlin, 2012/07/23 12:47:36]

None not empty string? We may want to test for the existence of a title, and
this would make it a bit ambiguous.

[cduvall, 2012/07/23 20:24:21]

I looked all throughout the internet for this, and I only found the super(Foo,
self).__init__() version. Thanks.

+    self._recent_tag = None
+    self._current = {}

[not at google - send to devlin, 2012/07/23 12:47:36]

current_what?

[cduvall, 2012/07/23 20:24:21]

Done.

+
+  def handle_starttag(self, tag, attrs):
+    id_ = ''
+    self._recent_tag = tag
+    for attr in attrs:
+      if attr[0] == 'id':
+        id_ = attr[1]
+    if tag == 'h2':
+      self._current = { 'link': id_, 'subheadings': [] }
+      self.toc.append(self._current)
+    elif tag == 'h3':
+      self._current = { 'link': id_ }
+      self.toc[-1]['subheadings'].append(self._current)
+
+  def handle_data(self, data):
+    if data.isspace():
+      return

[not at google - send to devlin, 2012/07/23 12:47:36]

why is this needed?

[cduvall, 2012/07/23 20:24:21]

Not needed anymore.

+    if self._recent_tag == 'h1':
+      self.page_title = data
+    elif self._recent_tag in ['h2', 'h3']:
+      self._current['title'] = data

[not at google - send to devlin, 2012/07/23 12:47:36]

Note that this won't handle cases like

<h2>This heading is <b>extra important</b></h2>

You might need to maintain a "header_tag_stack" and rather than accessing
recent_tag, access the top of that.

[cduvall, 2012/07/23 20:24:21]

I ended up not using a stack, but the new version handles tags in the headings.
It doesn't handle headings inside of headings, but that should never happen. I
can change it to use a stack instead if you would like.

[not at google - send to devlin, 2012/07/23 23:30:41]

sgtm. I realise I started micro-managing you a bit recently, will try to stop :)

[cduvall, 2012/07/23 23:58:33]

Np, I would much rather learn to do it the right way :)

+
 class IntroDataSource(object):
   """This class fetches the intros for a given API. From this intro, a table
   of contents dictionary is created, which contains the headings in the intro.
   """
   def __init__(self, cache_builder, base_paths):
     self._cache = cache_builder.build(self._MakeIntroDict)
     self._base_paths = base_paths
 
   def _MakeIntroDict(self, intro):
-    h1s = re.findall('<h1.*>(.+)</h1>', intro)
-    if len(h1s) > 0:
-      page_title = h1s[0]
-    else:
-      page_title = ''
-    headings = re.findall('<h([23]) id\="(.+)">(.+)</h[23]>', intro)
-    toc = []
-    for heading in headings:
-      level, link, title = heading
-      if level == '2':
-        toc.append({ 'link': link, 'title': title, 'subheadings': [] })
-      else:
-        toc[-1]['subheadings'].append({ 'link': link, 'title': title })
-    return { 'intro': Handlebar(intro), 'toc': toc , 'title': page_title }
+    parser = _IntroParser()
+    parser.init()
+    parser.feed(intro)
+    return {
+      'intro': Handlebar(intro),
+      'toc': parser.toc,
+      'title': parser.page_title
+    }
 
   def __getitem__(self, key):
     return self.get(key)
 
   def get(self, key):
     real_path = FormatKey(key)
     for base_path in self._base_paths:
       try:
         return self._cache.GetFromFile(base_path + '/' + real_path)
-      except Exception:
+      except Exception as e:
         pass
     return None