youtube_dl/extractor/justintv.py

   1 import json
   2 import os
   3 import re
   4 import xml.etree.ElementTree
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     ExtractorError,
   9     formatSeconds,
  10 )
  11
  12
  13 class JustinTVIE(InfoExtractor):
  14     """Information extractor for justin.tv and twitch.tv"""
  15     # TODO: One broadcast may be split into multiple videos. The key
  16     # 'broadcast_id' is the same for all parts, and 'broadcast_part'
  17     # starts at 1 and increases. Can we treat all parts as one video?
  18
  19     _VALID_URL = r"""(?x)^(?:http://)?(?:www\.)?(?:twitch|justin)\.tv/
  20         (?:
  21             (?P<channelid>[^/]+)|
  22             (?:(?:[^/]+)/b/(?P<videoid>[^/]+))|
  23             (?:(?:[^/]+)/c/(?P<chapterid>[^/]+))
  24         )
  25         /?(?:\#.*)?$
  26         """
  27     _JUSTIN_PAGE_LIMIT = 100
  28     IE_NAME = u'justin.tv'
  29     _TEST = {
  30         u'url': u'http://www.twitch.tv/thegamedevhub/b/296128360',
  31         u'file': u'296128360.flv',
  32         u'md5': u'ecaa8a790c22a40770901460af191c9a',
  33         u'info_dict': {
  34             u"upload_date": u"20110927",
  35             u"uploader_id": 25114803,
  36             u"uploader": u"thegamedevhub",
  37             u"title": u"Beginner Series - Scripting With Python Pt.1"
  38         }
  39     }
  40
  41     def report_download_page(self, channel, offset):
  42         """Report attempt to download a single page of videos."""
  43         self.to_screen(u'%s: Downloading video information from %d to %d' %
  44                 (channel, offset, offset + self._JUSTIN_PAGE_LIMIT))
  45
  46     # Return count of items, list of *valid* items
  47     def _parse_page(self, url, video_id):
  48         info_json = self._download_webpage(url, video_id,
  49                                            u'Downloading video info JSON',
  50                                            u'unable to download video info JSON')
  51
  52         response = json.loads(info_json)
  53         if type(response) != list:
  54             error_text = response.get('error', 'unknown error')
  55             raise ExtractorError(u'Justin.tv API: %s' % error_text)
  56         info = []
  57         for clip in response:
  58             video_url = clip['video_file_url']
  59             if video_url:
  60                 video_extension = os.path.splitext(video_url)[1][1:]
  61                 video_date = re.sub('-', '', clip['start_time'][:10])
  62                 video_uploader_id = clip.get('user_id', clip.get('channel_id'))
  63                 video_id = clip['id']
  64                 video_title = clip.get('title', video_id)
  65                 info.append({
  66                     'id': video_id,
  67                     'url': video_url,
  68                     'title': video_title,
  69                     'uploader': clip.get('channel_name', video_uploader_id),
  70                     'uploader_id': video_uploader_id,
  71                     'upload_date': video_date,
  72                     'ext': video_extension,
  73                 })
  74         return (len(response), info)
  75
  76     def _real_extract(self, url):
  77         mobj = re.match(self._VALID_URL, url)
  78         if mobj is None:
  79             raise ExtractorError(u'invalid URL: %s' % url)
  80
  81         api_base = 'http://api.justin.tv'
  82         paged = False
  83         if mobj.group('channelid'):
  84             paged = True
  85             video_id = mobj.group('channelid')
  86             api = api_base + '/channel/archives/%s.json' % video_id
  87         elif mobj.group('chapterid'):
  88             chapter_id = mobj.group('chapterid')
  89
  90             webpage = self._download_webpage(url, chapter_id)
  91             m = re.search(r'PP\.archive_id = "([0-9]+)";', webpage)
  92             if not m:
  93                 raise ExtractorError(u'Cannot find archive of a chapter')
  94             archive_id = m.group(1)
  95
  96             api = api_base + '/broadcast/by_chapter/%s.xml' % chapter_id
  97             chapter_info_xml = self._download_webpage(api, chapter_id,
  98                                              note=u'Downloading chapter information',
  99                                              errnote=u'Chapter information download failed')
 100             doc = xml.etree.ElementTree.fromstring(chapter_info_xml)
 101             for a in doc.findall('.//archive'):
 102                 if archive_id == a.find('./id').text:
 103                     break
 104             else:
 105                 raise ExtractorError(u'Could not find chapter in chapter information')
 106
 107             video_url = a.find('./video_file_url').text
 108             video_ext = video_url.rpartition('.')[2] or u'flv'
 109
 110             chapter_api_url = u'https://api.twitch.tv/kraken/videos/c' + chapter_id
 111             chapter_info_json = self._download_webpage(chapter_api_url, u'c' + chapter_id,
 112                                    note='Downloading chapter metadata',
 113                                    errnote='Download of chapter metadata failed')
 114             chapter_info = json.loads(chapter_info_json)
 115
 116             bracket_start = int(doc.find('.//bracket_start').text)
 117             bracket_end = int(doc.find('.//bracket_end').text)
 118
 119             # TODO determine start (and probably fix up file)
 120             #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
 121             #video_url += u'?start=' + TODO:start_timestamp
 122             # bracket_start is 13290, but we want 51670615
 123             self._downloader.report_warning(u'Chapter detected, but we can just download the whole file. '
 124                                             u'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
 125
 126             info = {
 127                 'id': u'c' + chapter_id,
 128                 'url': video_url,
 129                 'ext': video_ext,
 130                 'title': chapter_info['title'],
 131                 'thumbnail': chapter_info['preview'],
 132                 'description': chapter_info['description'],
 133                 'uploader': chapter_info['channel']['display_name'],
 134                 'uploader_id': chapter_info['channel']['name'],
 135             }
 136             return [info]
 137         else:
 138             video_id = mobj.group('videoid')
 139             api = api_base + '/broadcast/by_archive/%s.json' % video_id
 140
 141         self.report_extraction(video_id)
 142
 143         info = []
 144         offset = 0
 145         limit = self._JUSTIN_PAGE_LIMIT
 146         while True:
 147             if paged:
 148                 self.report_download_page(video_id, offset)
 149             page_url = api + ('?offset=%d&limit=%d' % (offset, limit))
 150             page_count, page_info = self._parse_page(page_url, video_id)
 151             info.extend(page_info)
 152             if not paged or page_count != limit:
 153                 break
 154             offset += limit
 155         return info