proboards-website-generator/generator.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

# 2015 by Sebastian Hugentobler <shugentobler@vanwa.ch>
# To the extent possible under law, the author(s) have dedicated all copyright
# and related and neighboring rights to this software to the public domain
# worldwide. This software is distributed without any warranty.
# See http://creativecommons.org/publicdomain/zero/1.0/ for a description
# of CC0.

import argparse
import json
import os
import re
import shutil

from datetime import datetime
from jinja2 import Environment, FileSystemLoader

css = """
body {
  margin: 2em auto;
  max-width: 50em;
  line-height: 1.6;
  font-size: 1.1em;
  color: #444; 
  padding: 0 1em;
}

h1,h2,h3 {
  line-height: 1.2;
}
"""


def fromunixtime(value):
    return datetime.fromtimestamp(value).strftime('%Y-%m-%d %H:%M:%S')


def url_replacer(value):
    return re.sub('[^a-zA-Z0-9\n\.]', '_', value)


def user_replacer(match):
    return '[url=../../user/' + url_replacer(match.group(2)) + '.html]' + match.group(2) + '[/url]'


def tohtml(value):
    value = value.replace('{{baseurl}}', 'static')
    value = value.replace('\n', '<br />')
    value = re.sub(r'\[url=\/user\/(.*?)\](.*?)\[\/url\]',
                   user_replacer, value)
    value = re.sub(r'\[url=(.*?)\](.*?)\[/url\]',
                   r'<a href="\1">\2</a>', value)
    value = re.sub(r'\[video\](.*?)\[/video\]', r'<a href="\1">\1</a>', value)
    value = re.sub(r'\[colour=(.*?)\](.*?)\[/colour\]',
                   r'<font color="\1">\2</font>', value)
    value = re.sub(r'\[b\](.*?)\[/b\]', r'<b>\1</b>', value)
    value = re.sub(r'\[i\](.*?)\[/i\]', r'<i>\1</i>', value)
    value = re.sub(r'\[u\](.*?)\[/u\]', r'<u>\1</u>', value)
    value = re.sub(r'\[img\](.*?)\[/img\]', r'<img src="\1">', value)

    for i in range(25):  # ugly hack but works good enough
        value = re.sub(r'\[quote=(.+?)\](.+)\[/quote\]', '<fieldset><legend>' + '<a href="../../user/' +
                       url_replacer(r'\1') + '.html">' + r'\1</a></legend>\2</fieldset>', value, count=1)

    for i in range(25):  # same here, shut up
        value = re.sub(r'\[quote\](.*?)\[/quote\]',
                       r'<fieldset>\1</fieldset>', value)

    return value


def write_render(rendered, name, outpath):
    if not os.path.exists(os.path.join(outpath, 'board', 'thread')):
        os.makedirs(os.path.join(outpath, 'board', 'thread'))

    if not os.path.exists(os.path.join(outpath, 'user')):
        os.makedirs(os.path.join(outpath, 'user'))

    with open(os.path.join(outpath, name), 'w') as f:
        f.write(rendered)


def find_unregistered_users(data):
    unregistered_users = []
    for board in data['boards']:
        for thread in board['threads']:
            for post in thread['posts']:
                if post['user']['name'] not in unregistered_users:
                    found = False
                    for user in data['users']:
                        if post['user']['name'] == user['name']:
                            found = True
                            break

                    if not found:
                        unregistered_users.append(post['user']['name'])

    return unregistered_users


def find_board_id(board):
    return board['link'].split('/')[-2]


def render_boards(boards, template_board, template_thread, outpath, title):
    for board in boards:
        rendered_board = template_board.render(
            board=board, title=title + ' - ' + board['title'])
        write_render(rendered_board, os.path.join('board', find_board_id(
            board) + '_' + url_replacer(board['title']) + '.html'), outpath)

        for thread in board['threads']:
            rendered_thread = template_thread.render(
                board=board, thread=thread, title=title + ' - ' + thread['title'])
            write_render(rendered_thread, os.path.join('board', 'thread', thread[
                         'id'] + '_' + url_replacer(thread['title']) + '.html'), outpath)

        render_boards(board['boards'], template_board,
                      template_thread, outpath, title)


def render(inputfile, staticpath, outpath, title):
    write_render(css, 'styles.css', outpath)

    with open(inputfile) as data_file:
        data = json.load(data_file)

        unregistered_users = find_unregistered_users(data)
        for unregistered_user in unregistered_users:
            data['users'].append(
                {'name': unregistered_user, 'registered': None})

        env = Environment(loader=FileSystemLoader('./templates'))
        env.filters['fromunixtime'] = fromunixtime
        env.filters['tohtml'] = tohtml
        env.filters['url_replacer'] = url_replacer
        env.filters['find_board_id'] = find_board_id

        template_users = env.get_template('users.html.j2')
        template_user = env.get_template('user.html.j2')
        template_boards = env.get_template('boards.html.j2')
        template_board = env.get_template('board.html.j2')
        template_thread = env.get_template('thread.html.j2')

        rendered_users = template_users.render(
            users=data['users'], title=title + ' - Users')
        rendered_boards = template_boards.render(
            boards=data['boards'], title=title + ' - Boards')

        write_render(rendered_users, 'users.html', outpath)
        write_render(rendered_boards, 'boards.html', outpath)

        shutil.rmtree(os.path.join(outpath, 'board',
                                   'thread', 'static'), ignore_errors=True)
        shutil.copytree(staticpath, os.path.join(
            outpath, 'board', 'thread', 'static'))

        for user in data['users']:
            rendered_user = template_user.render(
                user=user, title=title + ' - ' + user['name'])
            write_render(rendered_user, os.path.join(
                'user', url_replacer(user['name']) + '.html'), outpath)

        render_boards(data['boards'], template_board,
                      template_thread, outpath, title)

if __name__ == "__main__":
    parser = argparse.ArgumentParser(
        description='build a static website out of a proboard json dump')
    parser.add_argument('--data', default='board.json',
                        help='board data (json file)')
    parser.add_argument('--static', default='static',
                        help='path to the static files (images, attachments)')
    parser.add_argument('--out', default='rendered',
                        help='path where the website gets rendered to')
    parser.add_argument('--title', default='Proboard',
                        help='title for your pages')

    args = parser.parse_args()

    render(args.data, args.static, args.out, args.title)