#!/usr/bin/python3
# canonicalize html dirs to ease comaring them
#
# run as:
# ./tools/c10e-html html

import argparse
import glob
import re
import os
import sys

from bs4 import BeautifulSoup


def prettify(filename, parser='lxml', fixup=False):
    with open(filename, 'r') as doc:
        soup = BeautifulSoup(doc.read(), parser)
    with open(filename, 'w') as doc:
        html = soup.prettify()
        if fixup:
            # strip things that mkhtml2 is not producing to reduce the diff
            html = html.replace('a class="link" href', 'a href')
            html = html.replace(' target="_top"', '')
            html = html.replace('summary="Navigation header" ', '')
            html = html.replace("""   <a name="idx">
   </a>
""", '')
            html = re.sub("""  <div class="footer">
   <hr/>
   Generated by GTK-Doc V[.0-9]*
  </div>
""", '', html)
            html = re.sub(r'\s*<p>\s*</p>', '', html)
            html = re.sub(r'\s*<a name="id-[.0-9]+">\s*</a>', '', html)
            html = re.sub(r'\s*<div class="titlepage">\s*</div>', '', html)
            html = re.sub(r'\s*<meta content="DocBook[^>]*>', '', html)
            html = re.sub(r'\s*<meta content="GTK-Doc[^>]*>', '', html)
        doc.write(html)


def main(htmldir):
    for filename in glob.glob(os.path.join(htmldir, '*.devhelp2')):
        prettify(filename, parser='lxml-xml')
    for filename in glob.glob(os.path.join(htmldir, '*.html')):
        prettify(filename, fixup=True)


if __name__ == '__main__':
    parser = argparse.ArgumentParser(
        description='c10e-html - canonicalize html files for diffing')
    parser.add_argument('args', nargs='*',  help='HTML_DIR')

    options = parser.parse_args()
    if len(options.args) < 1:
        sys.exit('Too few arguments')

    main(options.args[0])