#!/usr/bin/python3
# SPDX-License-Identifier: LGPL-2.1-or-later

# Extracts translatable strings from HTML files in the following forms:
#
# <tag translate>String</tag>
# <tag translate context="value">String</tag>
#
# Note that some of the use of the translated may not support all the strings
# depending on the code actually using these strings to translate the HTML.


import argparse
import collections.abc
import json
import pathlib
from collections import defaultdict
from html.parser import HTMLParser

PO_HEADER = r"""msgid ""
msgstr ""
"Project-Id-Version: PACKAGE_VERSION\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=UTF-8\n"
"Content-Transfer-Encoding: 8bit\n"
"X-Generator: Cockpit html2po\n"
"""


class MyHTMLParser(HTMLParser):
    current_attr: tuple[str | None, str] | None = None
    filename: str

    def __init__(self, filename: str):
        self.strings = defaultdict[tuple[str | None, str], set[str]](set)
        self.filename = filename
        super().__init__()

    def handle_starttag(self, tag, attrs):
        translatable = False
        context = None

        # attrs => (key, value) where value needs to be split
        for (attr, value) in attrs:
            if attr == 'translate-context':
                context = value
            elif attr == 'context':
                context = value
            elif attr == 'translate' and (value is None or 'yes' in value.split(' ')):
                translatable = True

        if not translatable:
            return

        lineno = self.getpos()[0]
        filename = f'{self.filename}:{lineno}'
        self.current_attr = (context, filename)

    def handle_data(self, data: str):
        if self.current_attr and data:
            context, filename = self.current_attr
            self.strings[context, data].add(filename)
            self.current_attr = None


def main() -> None:
    parser = argparse.ArgumentParser(prog='html2po',
                                     description='Extracts translatable strings from HTML files')
    parser.add_argument('-d', '--directory', help='Base directory for input files')
    parser.add_argument('-o', '--output', help='Output file')
    parser.add_argument('files', nargs='+', help='One or more input files', type=pathlib.Path, metavar='FILE')

    args = parser.parse_args()
    strings = defaultdict[tuple[str | None, str], set[str]](set)

    files: collections.abc.Iterable[pathlib.Path] = args.files
    for file in files:
        # Qualify the filename if necessary
        full_path = args.directory / file if args.directory else file

        htmlparser = MyHTMLParser(str(file))
        with open(full_path, 'r') as fp:
            htmlparser.feed(fp.read())
            for ctx_msgid, filename in htmlparser.strings.items():
                strings[ctx_msgid].update(filename)

    with open(args.output, 'w') as fp:
        fp.write(PO_HEADER)
        for (context, msgid), filenames in strings.items():
            fp.write(f"\n#: {' '.join(filenames)}\n")
            # json.dumps() to escape any quotes in translation text or context
            if context:
                fp.write(f'msgctxt {json.dumps(context)}\n')

            fp.write(f'msgid {json.dumps(msgid)}\n')
            fp.write('msgstr ""\n')


if __name__ == "__main__":
    main()