X-Git-Url: http://git.cielonegro.org/gitweb.cgi?a=blobdiff_plain;ds=inline;f=youtube_dl%2Fextractor%2Ftwitch.py;h=20919774d0bcf65705e1f379a433667e2a06b7e9;hb=76dad392f5bd82493777d8efc35bcfccf70fafec;hp=df809ecfecb841a3d141c17571a3699fad6949a3;hpb=17b41a3337b5326fd2c8dec8aa1ab285225e1daa;p=youtube-dl.git
diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py
index df809ecfe..20919774d 100644
--- a/youtube_dl/extractor/twitch.py
+++ b/youtube_dl/extractor/twitch.py
@@ -7,13 +7,21 @@ import random
from .common import InfoExtractor
from ..compat import (
+ compat_parse_qs,
compat_str,
- compat_urllib_parse,
- compat_urllib_request,
+ compat_urllib_parse_urlencode,
+ compat_urllib_parse_urlparse,
+ compat_urlparse,
)
from ..utils import (
ExtractorError,
+ int_or_none,
+ js_to_json,
+ orderedSet,
+ parse_duration,
parse_iso8601,
+ sanitized_Request,
+ urlencode_postdata,
)
@@ -22,8 +30,7 @@ class TwitchBaseIE(InfoExtractor):
_API_BASE = 'https://api.twitch.tv'
_USHER_BASE = 'http://usher.twitch.tv'
- _LOGIN_URL = 'https://secure.twitch.tv/login'
- _LOGIN_POST_URL = 'https://passport.twitch.tv/authorize'
+ _LOGIN_URL = 'http://www.twitch.tv/login'
_NETRC_MACHINE = 'twitch'
def _handle_error(self, response):
@@ -43,7 +50,7 @@ class TwitchBaseIE(InfoExtractor):
for cookie in self._downloader.cookiejar:
if cookie.name == 'api_token':
headers['Twitch-Api-Token'] = cookie.value
- request = compat_urllib_request.Request(url, headers=headers)
+ request = sanitized_Request(url, headers=headers)
response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
self._handle_error(response)
return response
@@ -56,21 +63,28 @@ class TwitchBaseIE(InfoExtractor):
if username is None:
return
- login_page = self._download_webpage(
+ login_page, handle = self._download_webpage_handle(
self._LOGIN_URL, None, 'Downloading login page')
- login_form = dict(re.findall(
- r']+action=(["\'])(?P.+?)\1', login_page,
+ 'post url', default=redirect_url, group='url')
+
+ if not post_url.startswith('http'):
+ post_url = compat_urlparse.urljoin(redirect_url, post_url)
+
+ request = sanitized_Request(
+ post_url, urlencode_postdata(login_form))
+ request.add_header('Referer', redirect_url)
response = self._download_webpage(
request, None, 'Logging in as %s' % username)
@@ -131,14 +145,14 @@ class TwitchItemBaseIE(TwitchBaseIE):
def _extract_info(self, info):
return {
'id': info['_id'],
- 'title': info['title'],
- 'description': info['description'],
- 'duration': info['length'],
- 'thumbnail': info['preview'],
- 'uploader': info['channel']['display_name'],
- 'uploader_id': info['channel']['name'],
- 'timestamp': parse_iso8601(info['recorded_at']),
- 'view_count': info['views'],
+ 'title': info.get('title') or 'Untitled Broadcast',
+ 'description': info.get('description'),
+ 'duration': int_or_none(info.get('length')),
+ 'thumbnail': info.get('preview'),
+ 'uploader': info.get('channel', {}).get('display_name'),
+ 'uploader_id': info.get('channel', {}).get('name'),
+ 'timestamp': parse_iso8601(info.get('recorded_at')),
+ 'view_count': int_or_none(info.get('views')),
}
def _real_extract(self, url):
@@ -158,6 +172,7 @@ class TwitchVideoIE(TwitchItemBaseIE):
'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
},
'playlist_mincount': 12,
+ 'skip': 'HTTP Error 404: Not Found',
}
@@ -174,6 +189,7 @@ class TwitchChapterIE(TwitchItemBaseIE):
'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
},
'playlist_mincount': 3,
+ 'skip': 'HTTP Error 404: Not Found',
}, {
'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
'only_matching': True,
@@ -186,8 +202,8 @@ class TwitchVodIE(TwitchItemBaseIE):
_ITEM_TYPE = 'vod'
_ITEM_SHORTCUT = 'v'
- _TEST = {
- 'url': 'http://www.twitch.tv/riotgames/v/6528877',
+ _TESTS = [{
+ 'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
'info_dict': {
'id': 'v6528877',
'ext': 'mp4',
@@ -199,25 +215,62 @@ class TwitchVodIE(TwitchItemBaseIE):
'uploader': 'Riot Games',
'uploader_id': 'riotgames',
'view_count': int,
+ 'start_time': 310,
},
'params': {
# m3u8 download
'skip_download': True,
},
- }
+ }, {
+ # Untitled broadcast (title is None)
+ 'url': 'http://www.twitch.tv/belkao_o/v/11230755',
+ 'info_dict': {
+ 'id': 'v11230755',
+ 'ext': 'mp4',
+ 'title': 'Untitled Broadcast',
+ 'thumbnail': 're:^https?://.*\.jpg$',
+ 'duration': 1638,
+ 'timestamp': 1439746708,
+ 'upload_date': '20150816',
+ 'uploader': 'BelkAO_o',
+ 'uploader_id': 'belkao_o',
+ 'view_count': int,
+ },
+ 'params': {
+ # m3u8 download
+ 'skip_download': True,
+ },
+ }]
def _real_extract(self, url):
item_id = self._match_id(url)
+
info = self._download_info(self._ITEM_SHORTCUT, item_id)
access_token = self._download_json(
'%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
'Downloading %s access token' % self._ITEM_TYPE)
+
formats = self._extract_m3u8_formats(
- '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true'
- % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
- item_id, 'mp4')
+ '%s/vod/%s?%s' % (
+ self._USHER_BASE, item_id,
+ compat_urllib_parse_urlencode({
+ 'allow_source': 'true',
+ 'allow_audio_only': 'true',
+ 'allow_spectre': 'true',
+ 'player': 'twitchweb',
+ 'nauth': access_token['token'],
+ 'nauthsig': access_token['sig'],
+ })),
+ item_id, 'mp4', entry_protocol='m3u8_native')
+
self._prefer_source(formats)
info['formats'] = formats
+
+ parsed_url = compat_urllib_parse_urlparse(url)
+ query = compat_parse_qs(parsed_url.query)
+ if 't' in query:
+ info['start_time'] = parse_duration(query['t'][0])
+
return info
@@ -233,17 +286,37 @@ class TwitchPlaylistBaseIE(TwitchBaseIE):
entries = []
offset = 0
limit = self._PAGE_LIMIT
+ broken_paging_detected = False
+ counter_override = None
for counter in itertools.count(1):
response = self._download_json(
self._PLAYLIST_URL % (channel_id, offset, limit),
- channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
+ channel_id,
+ 'Downloading %s videos JSON page %s'
+ % (self._PLAYLIST_TYPE, counter_override or counter))
page_entries = self._extract_playlist_page(response)
if not page_entries:
break
+ total = int_or_none(response.get('_total'))
+ # Since the beginning of March 2016 twitch's paging mechanism
+ # is completely broken on the twitch side. It simply ignores
+ # a limit and returns the whole offset number of videos.
+ # Working around by just requesting all videos at once.
+ # Upd: pagination bug was fixed by twitch on 15.03.2016.
+ if not broken_paging_detected and total and len(page_entries) > limit:
+ self.report_warning(
+ 'Twitch pagination is broken on twitch side, requesting all videos at once',
+ channel_id)
+ broken_paging_detected = True
+ offset = total
+ counter_override = '(all at once)'
+ continue
entries.extend(page_entries)
+ if broken_paging_detected or total and len(page_entries) >= total:
+ break
offset += limit
return self.playlist_result(
- [self.url_result(entry) for entry in set(entries)],
+ [self.url_result(entry) for entry in orderedSet(entries)],
channel_id, channel_name)
def _extract_playlist_page(self, response):
@@ -285,36 +358,11 @@ class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
}
-class TwitchBookmarksIE(TwitchPlaylistBaseIE):
- IE_NAME = 'twitch:bookmarks'
- _VALID_URL = r'%s/(?P[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
- _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
- _PLAYLIST_TYPE = 'bookmarks'
-
- _TEST = {
- 'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
- 'info_dict': {
- 'id': 'ognos',
- 'title': 'Ognos',
- },
- 'playlist_mincount': 3,
- }
-
- def _extract_playlist_page(self, response):
- entries = []
- for bookmark in response.get('bookmarks', []):
- video = bookmark.get('video')
- if not video:
- continue
- entries.append(video['url'])
- return entries
-
-
class TwitchStreamIE(TwitchBaseIE):
IE_NAME = 'twitch:stream'
- _VALID_URL = r'%s/(?P[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
+ _VALID_URL = r'%s/(?P[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
- _TEST = {
+ _TESTS = [{
'url': 'http://www.twitch.tv/shroomztv',
'info_dict': {
'id': '12772022048',
@@ -333,7 +381,10 @@ class TwitchStreamIE(TwitchBaseIE):
# m3u8 download
'skip_download': True,
},
- }
+ }, {
+ 'url': 'http://www.twitch.tv/miracle_doto#profile-0',
+ 'only_matching': True,
+ }]
def _real_extract(self, url):
channel_id = self._match_id(url)
@@ -348,12 +399,19 @@ class TwitchStreamIE(TwitchBaseIE):
'http://www.twitch.tv/%s/profile' % channel_id,
'TwitchProfile', channel_id)
+ # Channel name may be typed if different case than the original channel name
+ # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
+ # an invalid m3u8 URL. Working around by use of original channel name from stream
+ # JSON and fallback to lowercase if it's not available.
+ channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
+
access_token = self._download_json(
'%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
'Downloading channel access token')
query = {
'allow_source': 'true',
+ 'allow_audio_only': 'true',
'p': random.randint(1000000, 10000000),
'player': 'twitchweb',
'segment_preference': '4',
@@ -362,7 +420,7 @@ class TwitchStreamIE(TwitchBaseIE):
}
formats = self._extract_m3u8_formats(
'%s/api/channel/hls/%s.m3u8?%s'
- % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
+ % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
channel_id, 'mp4')
self._prefer_source(formats)
@@ -397,3 +455,45 @@ class TwitchStreamIE(TwitchBaseIE):
'formats': formats,
'is_live': True,
}
+
+
+class TwitchClipsIE(InfoExtractor):
+ IE_NAME = 'twitch:clips'
+ _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P[^/?#&]+)'
+
+ _TEST = {
+ 'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
+ 'md5': '761769e1eafce0ffebfb4089cb3847cd',
+ 'info_dict': {
+ 'id': 'AggressiveCobraPoooound',
+ 'ext': 'mp4',
+ 'title': 'EA Play 2016 Live from the Novo Theatre',
+ 'thumbnail': 're:^https?://.*\.jpg',
+ 'creator': 'EA',
+ 'uploader': 'stereotype_',
+ 'uploader_id': 'stereotype_',
+ },
+ }
+
+ def _real_extract(self, url):
+ video_id = self._match_id(url)
+
+ webpage = self._download_webpage(url, video_id)
+
+ clip = self._parse_json(
+ self._search_regex(
+ r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
+ video_id, transform_source=js_to_json)
+
+ video_url = clip['clip_video_url']
+ title = clip['channel_title']
+
+ return {
+ 'id': video_id,
+ 'url': video_url,
+ 'title': title,
+ 'thumbnail': self._og_search_thumbnail(webpage),
+ 'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
+ 'uploader': clip.get('curator_login'),
+ 'uploader_id': clip.get('curator_display_name'),
+ }