new file mode 100755
--- /dev/null
+++ b/thp2epub.py
@@ -0,0 +1,283 @@
+#!/usr/bin/env python
+# -*- encoding: UTF-8 -*-
+##
+## Copyright © 2012 Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>
+##
+## This program is free software; you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published
+## by the Free Software Foundation; version 3 only.
+##
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+## GNU General Public License for more details.
+##
+
+
+from __future__ import unicode_literals
+
+from lxml import etree
+from lxml.builder import E
+from time import strptime, strftime
+import epub
+from argparse import ArgumentParser
+
+try:
+    from urllib.request import urlopen
+    from urllib.error import HTTPError
+except ImportError:
+    from urllib2 import urlopen
+    from urllib2 import HTTPError
+
+
+images_list = []
+mime_type = {
+    '.jpg': 'image/jpeg',
+    'jpeg': 'image/jpeg',
+    '.png': 'image/png',
+    '.gif': 'image/gif',
+    '.svg': 'image/svg+xml'
+}
+
+
+class Thread(object):
+    def __init__(self, op, replies):
+        self.op = op
+        self.replies = replies
+        self.title = op.title
+        self.author = op.author
+
+    def render(self, only_op):
+        html = E.html(
+            E.head(
+                E.title(self.title),
+                E.link(href='story.css', type='text/css', rel='stylesheet') #TODO: convert it to a PI.
+            ),
+            E.body(
+                E.h1(self.title)
+            ),
+            xmlns='http://www.w3.org/1999/xhtml'
+        )
+
+        body = html.find('body')
+        body.append(self.op.render(display_title=False))
+
+        for reply in self.replies:
+            # Remove user answers if not wanted.
+            if only_op and not reply.is_op(self.op):
+                continue
+
+            body.append(reply.render())
+
+        return html
+
+
+class Post(object):
+    def __init__(self, title, author, date, image, content):
+        self.title = title
+        self.author = author
+        self.date = date
+        self.image = image
+        self.content = content
+
+    def is_op(self, op):
+        return self.author.trip == op.author.trip
+
+    def render(self, display_title=True):
+        if display_title:
+            title = E.h2(self.title) if self.title and display_title else E.h2('⁂')
+        else:
+            title = ''
+
+        img = self.image.render() if self.image else ''
+
+        p = E.p()
+        for item in self.content:
+            #TODO: remove useless attributes like onclick.
+            p.append(item)
+
+        article = E.article(
+            title,
+            E.footer(
+                E.cite(self.author.render()),
+                ' ',
+                E.time(strftime('%y/%m/%d(%a)%H:%M', self.date), time=strftime('%y-%m-%dT%H:%M:%SZ', self.date))
+            ),
+            img,
+            p
+        )
+        return article
+
+
+class Image(object):
+    def __init__(self, name, filesize, size, url):
+        self.name = name
+        self.filesize = filesize
+        self.size = size
+        self.url = url
+
+    def render(self):
+        try:
+            url_file = urlopen(self.url)
+        except HTTPError:
+            return ''
+        with open(self.name, 'wb') as out:
+            out.write(url_file.read())
+        images_list.append(self.name)
+        return E.img(src=self.name, alt=self.name)
+
+
+class Author(object):
+    def __init__(self, name, trip, mail):
+        self.name = name
+        self.trip = trip
+        self.mail = mail
+
+    def render(self):
+        return '{}{}'.format(self.name, self.trip)
+
+
+def parse_post(root):
+    # We use the filesize element because it contains the image name.
+    label = root.find('label')
+    filesize = root.find('span[@class="filesize"]')
+    if filesize is not None:
+        a = filesize.getnext().getnext()
+        filesize = etree.tostring(filesize, method='text', encoding='UTF-8')
+        filesize = filesize.split()
+        name = filesize[7:-2]
+        name = b' '.join(name)
+        try:
+            name = name.decode('UTF-8')
+        except UnicodeDecodeError:
+            for i in range(-5, -42, -1):
+                char = name[i] if type(name[i]) is int else ord(name[i])
+                if char & 0xc0 == 0xc0:
+                    name = name[:i-1] + b'\xef\xbf\xbd' + name[-4:]
+                    break
+            name = name.decode('UTF-8')
+        filesize, size = filesize[3][1:], filesize[5]
+        if a.tag == 'a':
+            url = a.get('href')
+        else:
+            url = a.find('img').get('src')
+
+        image = Image(name, filesize, size, url)
+    else:
+        image = None
+
+    label = root.find('label')
+    title = label.find('span[@class="filetitle"]')
+    title = title.text.strip() if title is not None else None
+    name = label.find('span[@class="postername"]')
+    if name is not None:
+        last = name
+        a = name.find('a')
+        if a is not None:
+            mail = a.get('href')
+            name = a.text
+        else:
+            mail = ''
+            name = name.text
+    else:
+        mail = ''
+        name = ''
+
+    postertrip = label.find('span[@class="postertrip"]')
+    if postertrip is not None:
+        trip = postertrip.text
+        last = postertrip
+    else:
+        trip = ''
+
+    author = Author(name, trip, mail)
+
+    date = strptime(last.tail.strip(), '%y/%m/%d(%a)%H:%M')
+
+    blockquote = root.find('blockquote')
+    content = []
+    for item in blockquote:
+        if item is str and item.strip() == '':
+            continue
+        content.append(item)
+
+    return Post(title, author, date, image, content)
+
+
+def parse_thread(url):
+    tree = etree.parse(url, etree.HTMLParser())
+
+    root = tree.find('//form[@id="delform"]')
+    if root is None:
+        root = tree.find('//body')
+    op = parse_post(root)
+
+    replies = []
+    td = root.findall('.//td[@class="reply"]')
+    for reply in td:
+        replies.append(parse_post(reply))
+
+    return Thread(op, replies)
+
+
+def main(url, forum, only_op, threads):
+    threads_list = []
+    for thread in threads:
+        print('Rendering of thread №{}…'.format(thread))
+        t = parse_thread(url.format(forum, thread))
+        threads_list.append(t)
+
+        html = t.render(only_op)
+
+        # Use b mode as it allows us to directly dump UTF-8 data.
+        with open('{}.xhtml'.format(thread), 'wb') as f:
+            f.write(etree.tostring(html, pretty_print=True, xml_declaration=True, doctype='<!DOCTYPE html SYSTEM "/tmp/test.dtd">', encoding='UTF-8'))
+
+    with epub.open('story.epub', 'w') as book:
+        t = threads_list[0]
+
+        book.opf.metadata.add_title(t.title)
+        book.opf.metadata.add_creator(t.author.render())
+        book.opf.metadata.add_date(strftime('%y-%m-%dT%H:%M:%SZ'))
+        book.opf.metadata.add_language('en')
+
+        for thread in threads:
+            filename = '{}.xhtml'.format(thread)
+            manifest_item = epub.opf.ManifestItem(identifier='thread_{}'.format(thread),
+                                                  href=filename,
+                                                  media_type='application/xhtml+xml')
+            book.add_item(filename, manifest_item, True)
+
+        for image in images_list:
+            extension = image[-4:]
+            manifest_item = epub.opf.ManifestItem(identifier='image_{}'.format(image),
+                                                  href=image,
+                                                  media_type=mime_type[extension])
+            book.add_item(image, manifest_item, True)
+
+        manifest_item = epub.opf.ManifestItem(identifier='style',
+                                              href='story.css',
+                                              media_type='text/css')
+        book.add_item('story.css', manifest_item)
+
+        book.toc.title = t.title
+        nav_map = book.toc.nav_map
+        for thread in threads:
+            nav_point = epub.ncx.NavPoint()
+            nav_point.identifier = 'thread_%d' % thread
+            nav_point.add_label('Thread №%d' % thread)
+            nav_point.src = '%d.xhtml' % thread
+            nav_map.nav_point.append(nav_point)
+
+
+if __name__ == '__main__':
+    parser = ArgumentParser(description='Download and convert THP stories.')
+
+    parser.add_argument('threads', metavar='THREADS', nargs='+', type=int, help='List of the threads of the story.')
+    parser.add_argument('-u', '--url', metavar='URL', default='http://www.touhou-project.com/{}/res/{}.html', help='URL pattern from which the story will be downloaded, with the first {} as the forum, and the second as the thread.')
+    parser.add_argument('-f', '--forum', metavar='FORUM', default='sdm', help='The name of the forum (example: sdm, th, etc.).')
+    parser.add_argument('-o', '--only-op', action='store_true', help='Include only posts made by the original poster.')
+
+    args = parser.parse_args()
+
+    main(args.url, args.forum, args.only_op, args.threads)