|
|
- from __future__ import unicode_literals
-
- import datetime
- import re
-
- from .common import InfoExtractor
- from .subtitles import SubtitlesInfoExtractor
- from ..utils import (
- compat_str,
- compat_urllib_request,
-
- unescapeHTML,
- )
-
-
- class BlipTVIE(SubtitlesInfoExtractor):
- """Information extractor for blip.tv"""
-
- _VALID_URL = r'https?://(?:\w+\.)?blip\.tv/((.+/)|(play/)|(api\.swf#))(?P<presumptive_id>.+)$'
-
- _TESTS = [{
- 'url': 'http://blip.tv/cbr/cbr-exclusive-gotham-city-imposters-bats-vs-jokerz-short-3-5796352',
- 'md5': 'c6934ad0b6acf2bd920720ec888eb812',
- 'info_dict': {
- 'id': '5779306',
- 'ext': 'mov',
- 'upload_date': '20111205',
- 'description': 'md5:9bc31f227219cde65e47eeec8d2dc596',
- 'uploader': 'Comic Book Resources - CBR TV',
- 'title': 'CBR EXCLUSIVE: "Gotham City Imposters" Bats VS Jokerz Short 3',
- }
- }, {
- # https://github.com/rg3/youtube-dl/pull/2274
- 'note': 'Video with subtitles',
- 'url': 'http://blip.tv/play/h6Uag5OEVgI.html',
- 'md5': '309f9d25b820b086ca163ffac8031806',
- 'info_dict': {
- 'id': '6586561',
- 'ext': 'mp4',
- 'uploader': 'Red vs. Blue',
- 'description': 'One-Zero-One',
- 'upload_date': '20130614',
- 'title': 'Red vs. Blue Season 11 Episode 1',
- }
- }]
-
- def _real_extract(self, url):
- mobj = re.match(self._VALID_URL, url)
- presumptive_id = mobj.group('presumptive_id')
-
- # See https://github.com/rg3/youtube-dl/issues/857
- embed_mobj = re.match(r'https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)([a-zA-Z0-9]+)', url)
- if embed_mobj:
- info_url = 'http://blip.tv/play/%s.x?p=1' % embed_mobj.group(1)
- info_page = self._download_webpage(info_url, embed_mobj.group(1))
- video_id = self._search_regex(
- r'data-episode-id="([0-9]+)', info_page, 'video_id')
- return self.url_result('http://blip.tv/a/a-' + video_id, 'BlipTV')
-
- cchar = '&' if '?' in url else '?'
- json_url = url + cchar + 'skin=json&version=2&no_wrap=1'
- request = compat_urllib_request.Request(json_url)
- request.add_header('User-Agent', 'iTunes/10.6.1')
-
- json_data = self._download_json(request, video_id=presumptive_id)
-
- if 'Post' in json_data:
- data = json_data['Post']
- else:
- data = json_data
-
- video_id = compat_str(data['item_id'])
- upload_date = datetime.datetime.strptime(data['datestamp'], '%m-%d-%y %H:%M%p').strftime('%Y%m%d')
- subtitles = {}
- formats = []
- if 'additionalMedia' in data:
- for f in data['additionalMedia']:
- if f.get('file_type_srt') == 1:
- LANGS = {
- 'english': 'en',
- }
- lang = f['role'].rpartition('-')[-1].strip().lower()
- langcode = LANGS.get(lang, lang)
- subtitles[langcode] = f['url']
- continue
- if not int(f['media_width']): # filter m3u8
- continue
- formats.append({
- 'url': f['url'],
- 'format_id': f['role'],
- 'width': int(f['media_width']),
- 'height': int(f['media_height']),
- })
- else:
- formats.append({
- 'url': data['media']['url'],
- 'width': int(data['media']['width']),
- 'height': int(data['media']['height']),
- })
- self._sort_formats(formats)
-
- # subtitles
- video_subtitles = self.extract_subtitles(video_id, subtitles)
- if self._downloader.params.get('listsubtitles', False):
- self._list_available_subtitles(video_id, subtitles)
- return
-
- return {
- 'id': video_id,
- 'uploader': data['display_name'],
- 'upload_date': upload_date,
- 'title': data['title'],
- 'thumbnail': data['thumbnailUrl'],
- 'description': data['description'],
- 'user_agent': 'iTunes/10.6.1',
- 'formats': formats,
- 'subtitles': video_subtitles,
- }
-
- def _download_subtitle_url(self, sub_lang, url):
- # For some weird reason, blip.tv serves a video instead of subtitles
- # when we request with a common UA
- req = compat_urllib_request.Request(url)
- req.add_header('Youtubedl-user-agent', 'youtube-dl')
- return self._download_webpage(req, None, note=False)
-
-
- class BlipTVUserIE(InfoExtractor):
- _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?blip\.tv/)|bliptvuser:)([^/]+)/*$'
- _PAGE_SIZE = 12
- IE_NAME = 'blip.tv:user'
-
- def _real_extract(self, url):
- mobj = re.match(self._VALID_URL, url)
- username = mobj.group(1)
-
- page_base = 'http://m.blip.tv/pr/show_get_full_episode_list?users_id=%s&lite=0&esi=1'
-
- page = self._download_webpage(url, username, 'Downloading user page')
- mobj = re.search(r'data-users-id="([^"]+)"', page)
- page_base = page_base % mobj.group(1)
-
- # Download video ids using BlipTV Ajax calls. Result size per
- # query is limited (currently to 12 videos) so we need to query
- # page by page until there are no video ids - it means we got
- # all of them.
-
- video_ids = []
- pagenum = 1
-
- while True:
- url = page_base + "&page=" + str(pagenum)
- page = self._download_webpage(
- url, username, 'Downloading video ids from page %d' % pagenum)
-
- # Extract video identifiers
- ids_in_page = []
-
- for mobj in re.finditer(r'href="/([^"]+)"', page):
- if mobj.group(1) not in ids_in_page:
- ids_in_page.append(unescapeHTML(mobj.group(1)))
-
- video_ids.extend(ids_in_page)
-
- # A little optimization - if current page is not
- # "full", ie. does not contain PAGE_SIZE video ids then
- # we can assume that this page is the last one - there
- # are no more ids on further pages - no need to query
- # again.
-
- if len(ids_in_page) < self._PAGE_SIZE:
- break
-
- pagenum += 1
-
- urls = ['http://blip.tv/%s' % video_id for video_id in video_ids]
- url_entries = [self.url_result(vurl, 'BlipTV') for vurl in urls]
- return [self.playlist_result(url_entries, playlist_title=username)]
|