blob: 05c02c29de0c01902a4ff9637dc72fd762e73e4b [file] [log] [blame]
#!/usr/bin/env python
# Copyright (c) 2011 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.
'''This utility cleans up the html files as emitted by doxygen so
that they are suitable for publication on a Google documentation site.
import argparse
import glob
import os
import re
import shutil
import sys
from BeautifulSoup import BeautifulSoup, Tag
except (ImportError, NotImplementedError):
print ("This tool requires the BeautifulSoup package "
"Make sure that the file is either in this directory "
"or is available in your PYTHON_PATH")
def Trace(msg):
if Trace.verbose:
sys.stderr.write(str(msg) + '\n')
Trace.verbose = False
class HTMLFixer(object):
'''This class cleans up the html strings as produced by Doxygen
def __init__(self, html):
self.soup = BeautifulSoup(html)
def FixTableHeadings(self):
'''Fixes the doxygen table headings.
This includes:
- Using bare <h2> title row instead of row embedded in <tr><td> in table
- Putting the "name" attribute into the "id" attribute of the <tr> tag.
- Splitting up tables into multiple separate tables if a table
heading appears in the middle of a table.
For example, this html:
<tr><td colspan="2"><h2><a name="pub-attribs"></a>
Data Fields List</h2></td></tr>
would be converted to this:
<h2>Data Fields List</h2>
table_headers = []
for tag in self.soup.findAll('tr'):
if and and and['name']:
#tag['id'] =['name']
tag.string = = 'h2'
# reverse the list so that earlier tags don't delete later tags
# Split up tables that have multiple table header (th) rows
for tag in table_headers:
Trace("Header tag: %s is %s" % (, tag.string.strip()))
# Is this a heading in the middle of a table?
if tag.findPreviousSibling('tr') and == 'table':
Trace("Splitting Table named %s" % tag.string.strip())
table = tag.parent
table_parent = table.parent
table_index = table_parent.contents.index(table)
new_table = Tag(self.soup, name='table', attrs=table.attrs)
table_parent.insert(table_index + 1, new_table)
tag_index = table.contents.index(tag)
for index, row in enumerate(table.contents[tag_index:]):
new_table.insert(index, row)
# Now move the <h2> tag to be in front of the <table> tag
assert == 'table'
table = tag.parent
table_parent = table.parent
table_index = table_parent.contents.index(table)
table_parent.insert(table_index, tag)
def RemoveTopHeadings(self):
'''Removes <div> sections with a header, tabs, or navpath class attribute'''
header_tags = self.soup.findAll(
attrs={'class' : re.compile('^(header|tabs[0-9]*|navpath)$')})
[tag.extract() for tag in header_tags]
def RemoveVersionNumbers(self, html):
'''Horrible hack to strip _#_# from struct names.'''
return re.sub(r'(_\d_\d)(?=[": <])', '', html)
def FixAll(self):
html = str(self.soup)
html = self.RemoveVersionNumbers(html)
return html
def main(args):
"""Main entry for the doxy_cleanup utility
doxy_cleanup cleans up the html files generated by doxygen.
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument('-v', '--verbose', help='verbose output.',
options = parser.parse_args(args)
if options.verbose:
Trace.verbose = True
root_dir =
html_dir = os.path.join(root_dir, 'html')
# Doxygen puts all files in an 'html' directory.
# First, move all files from that directory to root_dir.
for filename in glob.glob(os.path.join(html_dir, '*')):
Trace('Moving %s -> %s' % (filename, root_dir))
shutil.move(filename, root_dir)
# Now remove the 'html' directory.
Trace('Removing %s' % html_dir)
# Then remove unneeded files.
for wildcard in FILES_TO_REMOVE:
Trace('Removing "%s":' % wildcard)
path = os.path.join(root_dir, wildcard)
for filename in glob.glob(path):
Trace(' Removing "%s"' % filename)
# Now, fix the HTML files we've kept.
Trace('Fixing HTML files...')
for root, _, files in os.walk(root_dir):
for filename in files:
if not os.path.splitext(filename)[1] == '.html':
Trace('Skipping %s' % filename)
filename = os.path.join(root, filename)
Trace('Processing "%s"...' % filename)
with open(filename) as f:
html =
fixer = HTMLFixer(html)
output = fixer.FixAll()
with open(filename, 'w') as f:
sys.stderr.write("Error while processing %s\n" % filename)
return 0
if __name__ == '__main__':
rtn = main(sys.argv[1:])
except KeyboardInterrupt:
sys.stderr.write('%s: interrupted\n' % os.path.basename(__file__))
rtn = 1