#!/usr/bin/env python3
import sys
import re
from ghtml import HTMLParser, HTMLParseError
from ggroff import Formatter

try:
    # Python 2 str - bytes version
    from StringIO import StringIO
except ImportError:
    # Python 3 str - unicode version
    from io import StringIO

entities = {"nbsp": " ", "bull": "*"}

# Remove ToC


def fix(content):
    if isinstance(content, tuple):
        tag, attrs, body = content
        if tag == "div" and ("class", "toc") in attrs:
            return None
        else:
            return (tag, attrs, fix(body))
    elif isinstance(content, list):
        return [fixed for item in content for fixed in [fix(item)] if fixed is not None]
    else:
        return content


def main():
    # parse HTML
    infile = sys.argv[1]
    inf = open(infile)
    p = HTMLParser(entities)
    for n, line in enumerate(inf):
        try:
            p.feed(line)
        except HTMLParseError as err:
            sys.stderr.write(
                "%s:%d:%d: Parse error: %s\n"
                % (infile, err.lineno, err.offset, err.msg)
            )
            sys.exit(1)
        except Exception as err:
            sys.stderr.write(
                "%s:%d:0: Error (%s): %s\n" % (infile, n + 1, repr(err), line)
            )
            sys.exit(1)
    p.close()
    inf.close()

    # generate groff
    sf = StringIO()
    f = Formatter(infile, sf)
    f.pp(fix(p.data))
    s = sf.getvalue()
    sf.close()

    # strip excess whitespace
    blank_re = re.compile("[ \t\n]*\n([ \t]*\n)*")
    s = blank_re.sub("\n", s)
    s = s.lstrip()

    # write groff
    with open(sys.argv[2], "wb") as outf:
        if sys.version_info.major >= 3:
            s = s.encode("UTF-8")
        outf.write(s)


if __name__ == "__main__":
    main()