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

   1 import json
   2 import os
   3 import re
   4
   5 from .common import InfoExtractor
   6 from ..utils import (
   7     ExtractorError,
   8     formatSeconds,
   9 )
  10
  11
  12 class JustinTVIE(InfoExtractor):
  13     """Information extractor for justin.tv and twitch.tv"""
  14     # TODO: One broadcast may be split into multiple videos. The key
  15     # 'broadcast_id' is the same for all parts, and 'broadcast_part'
  16     # starts at 1 and increases. Can we treat all parts as one video?
  17
  18     _VALID_URL = r"""(?x)^(?:http://)?(?:www\.)?(?:twitch|justin)\.tv/
  19         (?:
  20             (?P<channelid>[^/]+)|
  21             (?:(?:[^/]+)/b/(?P<videoid>[^/]+))|
  22             (?:(?:[^/]+)/c/(?P<chapterid>[^/]+))
  23         )
  24         /?(?:\#.*)?$
  25         """
  26     _JUSTIN_PAGE_LIMIT = 100
  27     IE_NAME = u'justin.tv'
  28     _TEST = {
  29         u'url': u'http://www.twitch.tv/thegamedevhub/b/296128360',
  30         u'file': u'296128360.flv',
  31         u'md5': u'ecaa8a790c22a40770901460af191c9a',
  32         u'info_dict': {
  33             u"upload_date": u"20110927",
  34             u"uploader_id": 25114803,
  35             u"uploader": u"thegamedevhub",
  36             u"title": u"Beginner Series - Scripting With Python Pt.1"
  37         }
  38     }
  39
  40     def report_download_page(self, channel, offset):
  41         """Report attempt to download a single page of videos."""
  42         self.to_screen(u'%s: Downloading video information from %d to %d' %
  43                 (channel, offset, offset + self._JUSTIN_PAGE_LIMIT))
  44
  45     # Return count of items, list of *valid* items
  46     def _parse_page(self, url, video_id):
  47         info_json = self._download_webpage(url, video_id,
  48                                            u'Downloading video info JSON',
  49                                            u'unable to download video info JSON')
  50
  51         response = json.loads(info_json)
  52         if type(response) != list:
  53             error_text = response.get('error', 'unknown error')
  54             raise ExtractorError(u'Justin.tv API: %s' % error_text)
  55         info = []
  56         for clip in response:
  57             video_url = clip['video_file_url']
  58             if video_url:
  59                 video_extension = os.path.splitext(video_url)[1][1:]
  60                 video_date = re.sub('-', '', clip['start_time'][:10])
  61                 video_uploader_id = clip.get('user_id', clip.get('channel_id'))
  62                 video_id = clip['id']
  63                 video_title = clip.get('title', video_id)
  64                 info.append({
  65                     'id': video_id,
  66                     'url': video_url,
  67                     'title': video_title,
  68                     'uploader': clip.get('channel_name', video_uploader_id),
  69                     'uploader_id': video_uploader_id,
  70                     'upload_date': video_date,
  71                     'ext': video_extension,
  72                 })
  73         return (len(response), info)
  74
  75     def _real_extract(self, url):
  76         mobj = re.match(self._VALID_URL, url)
  77         if mobj is None:
  78             raise ExtractorError(u'invalid URL: %s' % 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(u'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(api, chapter_id,
  97                                              note=u'Downloading chapter information',
  98                                              errnote=u'Chapter information download failed')
  99             for a in doc.findall('.//archive'):
 100                 if archive_id == a.find('./id').text:
 101                     break
 102             else:
 103                 raise ExtractorError(u'Could not find chapter in chapter information')
 104
 105             video_url = a.find('./video_file_url').text
 106             video_ext = video_url.rpartition('.')[2] or u'flv'
 107
 108             chapter_api_url = u'https://api.twitch.tv/kraken/videos/c' + chapter_id
 109             chapter_info_json = self._download_webpage(chapter_api_url, u'c' + chapter_id,
 110                                    note='Downloading chapter metadata',
 111                                    errnote='Download of chapter metadata failed')
 112             chapter_info = json.loads(chapter_info_json)
 113
 114             bracket_start = int(doc.find('.//bracket_start').text)
 115             bracket_end = int(doc.find('.//bracket_end').text)
 116
 117             # TODO determine start (and probably fix up file)
 118             #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
 119             #video_url += u'?start=' + TODO:start_timestamp
 120             # bracket_start is 13290, but we want 51670615
 121             self._downloader.report_warning(u'Chapter detected, but we can just download the whole file. '
 122                                             u'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
 123
 124             info = {
 125                 'id': u'c' + chapter_id,
 126                 'url': video_url,
 127                 'ext': video_ext,
 128                 'title': chapter_info['title'],
 129                 'thumbnail': chapter_info['preview'],
 130                 'description': chapter_info['description'],
 131                 'uploader': chapter_info['channel']['display_name'],
 132                 'uploader_id': chapter_info['channel']['name'],
 133             }
 134             return [info]
 135         else:
 136             video_id = mobj.group('videoid')
 137             api = api_base + '/broadcast/by_archive/%s.json' % video_id
 138
 139         self.report_extraction(video_id)
 140
 141         info = []
 142         offset = 0
 143         limit = self._JUSTIN_PAGE_LIMIT
 144         while True:
 145             if paged:
 146                 self.report_download_page(video_id, offset)
 147             page_url = api + ('?offset=%d&limit=%d' % (offset, limit))
 148             page_count, page_info = self._parse_page(page_url, video_id)
 149             info.extend(page_info)
 150             if not paged or page_count != limit:
 151                 break
 152             offset += limit
 153         return info