Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/justintv.py

   1 from __future__ import unicode_literals
   2
   3 import itertools
   4 import json
   5 import os
   6 import re
   7
   8 from .common import InfoExtractor
   9 from ..utils import (
  10     compat_str,
  11     ExtractorError,
  12     formatSeconds,
  13 )
  14
  15
  16 class JustinTVIE(InfoExtractor):
  17     """Information extractor for justin.tv and twitch.tv"""
  18     # TODO: One broadcast may be split into multiple videos. The key
  19     # 'broadcast_id' is the same for all parts, and 'broadcast_part'
  20     # starts at 1 and increases. Can we treat all parts as one video?
  21
  22     _VALID_URL = r"""(?x)^(?:http://)?(?:www\.)?(?:twitch|justin)\.tv/
  23         (?:
  24             (?P<channelid>[^/]+)|
  25             (?:(?:[^/]+)/b/(?P<videoid>[^/]+))|
  26             (?:(?:[^/]+)/c/(?P<chapterid>[^/]+))
  27         )
  28         /?(?:\#.*)?$
  29         """
  30     _JUSTIN_PAGE_LIMIT = 100
  31     IE_NAME = 'justin.tv'
  32     IE_DESC = 'justin.tv and twitch.tv'
  33     _TEST = {
  34         'url': 'http://www.twitch.tv/thegamedevhub/b/296128360',
  35         'md5': 'ecaa8a790c22a40770901460af191c9a',
  36         'info_dict': {
  37             'id': '296128360',
  38             'ext': 'flv',
  39             'upload_date': '20110927',
  40             'uploader_id': 25114803,
  41             'uploader': 'thegamedevhub',
  42             'title': 'Beginner Series - Scripting With Python Pt.1'
  43         }
  44     }
  45
  46     # Return count of items, list of *valid* items
  47     def _parse_page(self, url, video_id, counter):
  48         info_json = self._download_webpage(
  49             url, video_id,
  50             'Downloading video info JSON on page %d' % counter,
  51             'Unable to download video info JSON %d' % counter)
  52
  53         response = json.loads(info_json)
  54         if type(response) != list:
  55             error_text = response.get('error', 'unknown error')
  56             raise ExtractorError('Justin.tv API: %s' % error_text)
  57         info = []
  58         for clip in response:
  59             video_url = clip['video_file_url']
  60             if video_url:
  61                 video_extension = os.path.splitext(video_url)[1][1:]
  62                 video_date = re.sub('-', '', clip['start_time'][:10])
  63                 video_uploader_id = clip.get('user_id', clip.get('channel_id'))
  64                 video_id = clip['id']
  65                 video_title = clip.get('title', video_id)
  66                 info.append({
  67                     'id': compat_str(video_id),
  68                     'url': video_url,
  69                     'title': video_title,
  70                     'uploader': clip.get('channel_name', video_uploader_id),
  71                     'uploader_id': video_uploader_id,
  72                     'upload_date': video_date,
  73                     'ext': video_extension,
  74                 })
  75         return (len(response), info)
  76
  77     def _real_extract(self, url):
  78         mobj = re.match(self._VALID_URL, url)
  79
  80         api_base = 'http://api.justin.tv'
  81         paged = False
  82         if mobj.group('channelid'):
  83             paged = True
  84             video_id = mobj.group('channelid')
  85             api = api_base + '/channel/archives/%s.json' % video_id
  86         elif mobj.group('chapterid'):
  87             chapter_id = mobj.group('chapterid')
  88
  89             webpage = self._download_webpage(url, chapter_id)
  90             m = re.search(r'PP\.archive_id = "([0-9]+)";', webpage)
  91             if not m:
  92                 raise ExtractorError('Cannot find archive of a chapter')
  93             archive_id = m.group(1)
  94
  95             api = api_base + '/broadcast/by_chapter/%s.xml' % chapter_id
  96             doc = self._download_xml(
  97                 api, chapter_id,
  98                 note='Downloading chapter information',
  99                 errnote='Chapter information download failed')
 100             for a in doc.findall('.//archive'):
 101                 if archive_id == a.find('./id').text:
 102                     break
 103             else:
 104                 raise ExtractorError('Could not find chapter in chapter information')
 105
 106             video_url = a.find('./video_file_url').text
 107             video_ext = video_url.rpartition('.')[2] or 'flv'
 108
 109             chapter_api_url = 'https://api.twitch.tv/kraken/videos/c' + chapter_id
 110             chapter_info = self._download_json(
 111                 chapter_api_url, 'c' + chapter_id,
 112                 note='Downloading chapter metadata',
 113                 errnote='Download of chapter metadata failed')
 114
 115             bracket_start = int(doc.find('.//bracket_start').text)
 116             bracket_end = int(doc.find('.//bracket_end').text)
 117
 118             # TODO determine start (and probably fix up file)
 119             #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
 120             #video_url += '?start=' + TODO:start_timestamp
 121             # bracket_start is 13290, but we want 51670615
 122             self._downloader.report_warning('Chapter detected, but we can just download the whole file. '
 123                                             'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
 124
 125             info = {
 126                 'id': 'c' + chapter_id,
 127                 'url': video_url,
 128                 'ext': video_ext,
 129                 'title': chapter_info['title'],
 130                 'thumbnail': chapter_info['preview'],
 131                 'description': chapter_info['description'],
 132                 'uploader': chapter_info['channel']['display_name'],
 133                 'uploader_id': chapter_info['channel']['name'],
 134             }
 135             return info
 136         else:
 137             video_id = mobj.group('videoid')
 138             api = api_base + '/broadcast/by_archive/%s.json' % video_id
 139
 140         entries = []
 141         offset = 0
 142         limit = self._JUSTIN_PAGE_LIMIT
 143         for counter in itertools.count(1):
 144             page_url = api + ('?offset=%d&limit=%d' % (offset, limit))
 145             page_count, page_info = self._parse_page(
 146                 page_url, video_id, counter)
 147             entries.extend(page_info)
 148             if not paged or page_count != limit:
 149                 break
 150             offset += limit
 151         return {
 152             '_type': 'playlist',
 153             'id': video_id,
 154             'entries': entries,
 155         }