tools/c10e-html.py

   1 #!/usr/bin/python3
   2 # canonicalize html dirs to ease comaring them
   3 #
   4 # run as:
   5 # ./tools/c10e-html html
   6
   7 import argparse
   8 import glob
   9 import os
  10 import sys
  11
  12 from bs4 import BeautifulSoup
  13
  14
  15 def prettify(filename):
  16     with open(filename, 'r') as doc:
  17         soup = BeautifulSoup(doc.read(), 'lxml')
  18     with open(filename, 'w') as doc:
  19         doc.write(soup.prettify())
  20
  21
  22 def main(htmldir):
  23     for filename in glob.glob(os.path.join(htmldir, '*.devhelp2')):
  24         prettify(filename)
  25     for filename in glob.glob(os.path.join(htmldir, '*.html')):
  26         prettify(filename)
  27
  28
  29 if __name__ == '__main__':
  30     parser = argparse.ArgumentParser(
  31         description='c10e-html - canonicalize html files for diffing')
  32     parser.add_argument('args', nargs='*',  help='HTML_DIR')
  33
  34     options = parser.parse_args()
  35     if len(options.args) < 1:
  36         sys.exit('Too few arguments')
  37
  38     main(options.args[0])