_ Git - youtube-dl/blob - 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
  30     def report_download_page(self, channel, offset):
  31         """Report attempt to download a single page of videos."""
  32         self.to_screen(u'%s: Downloading video information from %d to %d' %
  33                 (channel, offset, offset + self._JUSTIN_PAGE_LIMIT))
  34
  35     # Return count of items, list of *valid* items
  36     def _parse_page(self, url, video_id):
  37         info_json = self._download_webpage(url, video_id,
  38                                            u'Downloading video info JSON',
  39                                            u'unable to download video info JSON')
  40
  41         response = json.loads(info_json)
  42         if type(response) != list:
  43             error_text = response.get('error', 'unknown error')
  44             raise ExtractorError(u'Justin.tv API: %s' % error_text)
  45         info = []
  46         for clip in response:
  47             video_url = clip['video_file_url']
  48             if video_url:
  49                 video_extension = os.path.splitext(video_url)[1][1:]
  50                 video_date = re.sub('-', '', clip['start_time'][:10])
  51                 video_uploader_id = clip.get('user_id', clip.get('channel_id'))
  52                 video_id = clip['id']
  53                 video_title = clip.get('title', video_id)
  54                 info.append({
  55                     'id': video_id,
  56                     'url': video_url,
  57                     'title': video_title,
  58                     'uploader': clip.get('channel_name', video_uploader_id),
  59                     'uploader_id': video_uploader_id,
  60                     'upload_date': video_date,
  61                     'ext': video_extension,
  62                 })
  63         return (len(response), info)
  64
  65     def _real_extract(self, url):
  66         mobj = re.match(self._VALID_URL, url)
  67         if mobj is None:
  68             raise ExtractorError(u'invalid URL: %s' % url)
  69
  70         api_base = 'http://api.justin.tv'
  71         paged = False
  72         if mobj.group('channelid'):
  73             paged = True
  74             video_id = mobj.group('channelid')
  75             api = api_base + '/channel/archives/%s.json' % video_id
  76         elif mobj.group('chapterid'):
  77             chapter_id = mobj.group('chapterid')
  78
  79             webpage = self._download_webpage(url, chapter_id)
  80             m = re.search(r'PP\.archive_id = "([0-9]+)";', webpage)
  81             if not m:
  82                 raise ExtractorError(u'Cannot find archive of a chapter')
  83             archive_id = m.group(1)
  84
  85             api = api_base + '/broadcast/by_chapter/%s.xml' % chapter_id
  86             chapter_info_xml = self._download_webpage(api, chapter_id,
  87                                              note=u'Downloading chapter information',
  88                                              errnote=u'Chapter information download failed')
  89             doc = xml.etree.ElementTree.fromstring(chapter_info_xml)
  90             for a in doc.findall('.//archive'):
  91                 if archive_id == a.find('./id').text:
  92                     break
  93             else:
  94                 raise ExtractorError(u'Could not find chapter in chapter information')
  95
  96             video_url = a.find('./video_file_url').text
  97             video_ext = video_url.rpartition('.')[2] or u'flv'
  98
  99             chapter_api_url = u'https://api.twitch.tv/kraken/videos/c' + chapter_id
 100             chapter_info_json = self._download_webpage(chapter_api_url, u'c' + chapter_id,
 101                                    note='Downloading chapter metadata',
 102                                    errnote='Download of chapter metadata failed')
 103             chapter_info = json.loads(chapter_info_json)
 104
 105             bracket_start = int(doc.find('.//bracket_start').text)
 106             bracket_end = int(doc.find('.//bracket_end').text)
 107
 108             # TODO determine start (and probably fix up file)
 109             #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
 110             #video_url += u'?start=' + TODO:start_timestamp
 111             # bracket_start is 13290, but we want 51670615
 112             self._downloader.report_warning(u'Chapter detected, but we can just download the whole file. '
 113                                             u'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
 114
 115             info = {
 116                 'id': u'c' + chapter_id,
 117                 'url': video_url,
 118                 'ext': video_ext,
 119                 'title': chapter_info['title'],
 120                 'thumbnail': chapter_info['preview'],
 121                 'description': chapter_info['description'],
 122                 'uploader': chapter_info['channel']['display_name'],
 123                 'uploader_id': chapter_info['channel']['name'],
 124             }
 125             return [info]
 126         else:
 127             video_id = mobj.group('videoid')
 128             api = api_base + '/broadcast/by_archive/%s.json' % video_id
 129
 130         self.report_extraction(video_id)
 131
 132         info = []
 133         offset = 0
 134         limit = self._JUSTIN_PAGE_LIMIT
 135         while True:
 136             if paged:
 137                 self.report_download_page(video_id, offset)
 138             page_url = api + ('?offset=%d&limit=%d' % (offset, limit))
 139             page_count, page_info = self._parse_page(page_url, video_id)
 140             info.extend(page_info)
 141             if not paged or page_count != limit:
 142                 break
 143             offset += limit
 144         return info