#!/usr/bin/env python3

# Copyright 2013-2015 Endless Mobile, Inc.

import argparse
import io
import os.path
import re
import sys
from bs4 import BeautifulSoup
from html.parser import HTMLParser

def normalize_string(string):
    return re.sub(r'\s+', ' ', string.strip())


# Parser that adds line numbers to the HTML strings that need translating
class TranslatableHTMLParser(HTMLParser):
    def __init__(self, translatable_strings):
        super().__init__()
        self.all_translatable_data = []
        self._comments_with_line_numbers = []
        self._translatable_strings = set(translatable_strings)

    def handle_data(self, data):
        if data not in self._translatable_strings:
            return

        code_line = self.getpos()[0]
        optional_comment = None

        if self._comments_with_line_numbers:
            # Determine if comment should be included
            most_recent_comment = self._comments_with_line_numbers[-1]
            comment_string, comment_line = most_recent_comment

            # Comment takes up at least one line by default (hence the +1)
            comment_length = len(re.findall(r'\n', comment_string)) + 1

            # If the comment immediately preceded this string, include it
            if comment_line + comment_length == code_line:
                optional_comment = ' '.join(comment_string.split())

        self.all_translatable_data.append((normalize_string(data), code_line, optional_comment))

    def handle_comment(self, comment):
        self._comments_with_line_numbers.append((comment, self.getpos()[0]))

# Ensure stdout is UTF-8
default_out = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')

parser = argparse.ArgumentParser(description='Extract translatable strings ' +
    'from HTML files. This is xgettext for HTML.')
parser.add_argument('input_file', type=str,
    help='Input file to scan')
parser.add_argument('top_srcdir', type=str, nargs='?', default='.',
    help='Top-level source directory (for printing correct #line directives)')
parser.add_argument('-o', '--output', default=default_out,
    type=argparse.FileType('w', encoding='utf-8'),
    help='File to write (default: stdout)')
args = parser.parse_args()

# Path from current directory to top-level app directory
html_file = args.input_file
top_dir = args.top_srcdir
final_path = os.path.relpath(html_file, top_dir)
out_file = args.output

# Create the BeautifulSoup HTML-parsing object
with open(html_file, encoding='utf-8') as f:
    page = f.read()
soup = BeautifulSoup(page)

# Extract all translatable strings from that HTML
translatable_divs = soup.find_all(attrs={'name': 'translatable'})
translatable_strings = map(lambda div: div.contents[0], translatable_divs)

# Find the line numbers for those strings
parser = TranslatableHTMLParser(translatable_strings)
parser.feed(page)

# Write out all info about the translatable strings found in this file
for string, line_num, optional_comment in parser.all_translatable_data:
    out_file.write('#line {line} "{path}"\n'.format(line=line_num, path=final_path))
    if optional_comment:
        out_file.write('// {}\n'.format(optional_comment))
    out_file.write('_("{string}");\n'.format(string=string))