mirror of
https://github.com/blackjack4494/yt-dlc.git
synced 2024-11-19 17:32:39 +01:00
Minor changes to make it easier to merge
This commit is contained in:
parent
70c5802b5d
commit
70d5c17b08
@ -1153,7 +1153,7 @@
|
||||
- **YourPorn**
|
||||
- **YourUpload**
|
||||
- **youtube**: YouTube.com
|
||||
- **youtube:favorites**: YouTube.com liked videos, ":ytfav" or "LL" for short (requires authentication)
|
||||
- **youtube:favorites**: YouTube.com liked videos, ":ytfav" for short (requires authentication)
|
||||
- **youtube:history**: Youtube watch history, ":ythistory" for short (requires authentication)
|
||||
- **youtube:playlist**: YouTube.com playlists
|
||||
- **youtube:recommended**: YouTube.com recommended videos, ":ytrec" for short (requires authentication)
|
||||
@ -1162,7 +1162,7 @@
|
||||
- **youtube:search_url**: YouTube.com search URLs
|
||||
- **youtube:subscriptions**: YouTube.com subscriptions feed, ":ytsubs" for short (requires authentication)
|
||||
- **youtube:tab**: YouTube.com tab
|
||||
- **youtube:watchlater**: Youtube watch later list, ":ytwatchlater" or "WL" for short (requires authentication)
|
||||
- **youtube:watchlater**: Youtube watch later list, ":ytwatchlater" for short (requires authentication)
|
||||
- **YoutubeYtUser**: YouTube.com user videos, URL or "ytuser" keyword
|
||||
- **Zapiks**
|
||||
- **Zaq1**
|
||||
|
@ -72,7 +72,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
|
||||
# If True it will raise an error if no login info is provided
|
||||
_LOGIN_REQUIRED = False
|
||||
|
||||
_PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM)'
|
||||
_PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM|WL|LL|LM)'
|
||||
|
||||
_YOUTUBE_CLIENT_HEADERS = {
|
||||
'x-youtube-client-name': '1',
|
||||
@ -1328,7 +1328,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
|
||||
# https://github.com/ytdl-org/youtube-dl/pull/7599)
|
||||
r';ytplayer\.config\s*=\s*({.+?});ytplayer',
|
||||
r';ytplayer\.config\s*=\s*({.+?});',
|
||||
r'ytInitialPlayerResponse\s*=\s*({.+?});var meta' # Needed???
|
||||
)
|
||||
config = self._search_regex(
|
||||
patterns, webpage, 'ytplayer.config', default=None)
|
||||
@ -2532,11 +2531,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
|
||||
|
||||
class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
IE_DESC = 'YouTube.com tab'
|
||||
_VALID_URL = (r'''(?x)
|
||||
https?://(?:\w+\.)?(?:youtube(?:kids)?\.com|invidio\.us)/(?:
|
||||
(?!(%s)([/#?]|$))|channel/|c/|user/|
|
||||
(?P<not_channel>playlist|watch)/?\?.*?\blist=)
|
||||
(?P<id>[^/?#&]+)''') % YoutubeBaseInfoExtractor._RESERVED_NAMES
|
||||
_VALID_URL = r'''(?x)
|
||||
https?://
|
||||
(?:\w+\.)?
|
||||
(?:
|
||||
youtube(?:kids)?\.com|
|
||||
invidio\.us
|
||||
)/
|
||||
(?:
|
||||
(?:channel|c|user)/|
|
||||
(?P<not_channel>
|
||||
(?:playlist|watch)\?.*?\blist=
|
||||
)|
|
||||
(?!(%s)([/#?]|$)) # Direct URLs
|
||||
)
|
||||
(?P<id>[^/?\#&]+)
|
||||
''' % YoutubeBaseInfoExtractor._RESERVED_NAMES
|
||||
IE_NAME = 'youtube:tab'
|
||||
|
||||
_TESTS = [{
|
||||
@ -2910,7 +2920,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
|
||||
def _rich_entries(self, rich_grid_renderer):
|
||||
renderer = try_get(
|
||||
rich_grid_renderer, lambda x: x['content']['videoRenderer'], dict)
|
||||
rich_grid_renderer, lambda x: x['content']['videoRenderer'], dict) or {}
|
||||
video_id = renderer.get('videoId')
|
||||
if not video_id:
|
||||
return
|
||||
@ -3008,14 +3018,14 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
|
||||
def _entries(self, tab, identity_token):
|
||||
|
||||
def extract_entries(parent_renderer):
|
||||
slr_contents = try_get(parent_renderer, lambda x: x['contents'], list) or []
|
||||
for slr_content in slr_contents:
|
||||
if not isinstance(slr_content, dict):
|
||||
def extract_entries(parent_renderer): # this needs to called again for continuation to work with feeds
|
||||
contents = try_get(parent_renderer, lambda x: x['contents'], list) or []
|
||||
for content in contents:
|
||||
if not isinstance(content, dict):
|
||||
continue
|
||||
is_renderer = try_get(slr_content, lambda x: x['itemSectionRenderer'], dict)
|
||||
is_renderer = try_get(content, lambda x: x['itemSectionRenderer'], dict)
|
||||
if not is_renderer:
|
||||
renderer = slr_content.get('richItemRenderer')
|
||||
renderer = content.get('richItemRenderer')
|
||||
if renderer:
|
||||
for entry in self._rich_entries(renderer):
|
||||
yield entry
|
||||
@ -3041,7 +3051,6 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
if renderer:
|
||||
for entry in self._shelf_entries(renderer):
|
||||
yield entry
|
||||
continuation_list[0] = self._extract_continuation(parent_renderer)
|
||||
continue
|
||||
renderer = isr_content.get('backstagePostThreadRenderer')
|
||||
if renderer:
|
||||
@ -3054,8 +3063,10 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
entry = self._video_entry(renderer)
|
||||
if entry:
|
||||
yield entry
|
||||
|
||||
if not continuation_list[0]:
|
||||
continuation_list[0] = self._extract_continuation(is_renderer)
|
||||
|
||||
if not continuation_list[0]:
|
||||
continuation_list[0] = self._extract_continuation(parent_renderer)
|
||||
|
||||
@ -3063,7 +3074,6 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
parent_renderer = (
|
||||
try_get(tab, lambda x: x['sectionListRenderer'], dict)
|
||||
or try_get(tab, lambda x: x['richGridRenderer'], dict) or {})
|
||||
if parent_renderer:
|
||||
for entry in extract_entries(parent_renderer):
|
||||
yield entry
|
||||
continuation = continuation_list[0]
|
||||
@ -3078,8 +3088,6 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
for page_num in itertools.count(1):
|
||||
if not continuation:
|
||||
break
|
||||
if hasattr(self, '_MAX_PAGES') and page_num > self._MAX_PAGES:
|
||||
break
|
||||
browse = self._download_json(
|
||||
'https://www.youtube.com/browse_ajax', None,
|
||||
'Downloading page %d' % page_num,
|
||||
@ -3111,7 +3119,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
yield entry
|
||||
continuation = self._extract_continuation(continuation_renderer)
|
||||
continue
|
||||
continuation_renderer = continuation_contents.get('sectionListContinuation')
|
||||
continuation_renderer = continuation_contents.get('sectionListContinuation') # for feeds
|
||||
if continuation_renderer:
|
||||
continuation_list = [None]
|
||||
for entry in extract_entries(continuation_renderer):
|
||||
@ -3125,19 +3133,13 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
continuation_item = continuation_items[0]
|
||||
if not isinstance(continuation_item, dict):
|
||||
continue
|
||||
renderer = continuation_item.get('playlistVideoRenderer')
|
||||
renderer = continuation_item.get('playlistVideoRenderer') or continuation_item.get('itemSectionRenderer')
|
||||
if renderer:
|
||||
video_list_renderer = {'contents': continuation_items}
|
||||
for entry in self._playlist_entries(video_list_renderer):
|
||||
yield entry
|
||||
continuation = self._extract_continuation(video_list_renderer)
|
||||
continue
|
||||
renderer = continuation_item.get('itemSectionRenderer')
|
||||
if renderer:
|
||||
for entry in self._itemSection_entries(renderer):
|
||||
yield entry
|
||||
continuation = self._extract_continuation({'contents': continuation_items})
|
||||
continue
|
||||
break
|
||||
|
||||
@staticmethod
|
||||
@ -3175,7 +3177,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
selected_tab = self._extract_selected_tab(tabs)
|
||||
renderer = try_get(
|
||||
data, lambda x: x['metadata']['channelMetadataRenderer'], dict)
|
||||
playlist_id = None
|
||||
playlist_id = title = description = None
|
||||
if renderer:
|
||||
channel_title = renderer.get('title') or item_id
|
||||
tab_title = selected_tab.get('title')
|
||||
@ -3191,7 +3193,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
description = None
|
||||
playlist_id = item_id
|
||||
if playlist_id is None:
|
||||
return None
|
||||
playlist_id = item_id
|
||||
if title is None:
|
||||
title = "Youtube " + playlist_id.title()
|
||||
playlist = self.playlist_result(
|
||||
self._entries(selected_tab['content'], identity_token),
|
||||
playlist_id=playlist_id, playlist_title=title,
|
||||
@ -3212,7 +3216,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
|
||||
url = compat_urlparse.urlunparse(
|
||||
compat_urlparse.urlparse(url)._replace(netloc='www.youtube.com'))
|
||||
is_home = re.match(r'(?P<pre>%s)(?P<post>/?(?![^#?]).*$)' % self._VALID_URL, url)
|
||||
if is_home is not None and is_home.group('not_channel') is None:
|
||||
if is_home is not None and is_home.group('not_channel') is None and item_id != 'feed':
|
||||
self._downloader.report_warning(
|
||||
'A channel/user page was given. All the channel\'s videos will be downloaded. '
|
||||
'To download only the videos in the home page, add a "/home" to the URL')
|
||||
@ -3365,6 +3369,25 @@ class YoutubeYtUserIE(InfoExtractor):
|
||||
ie=YoutubeTabIE.ie_key(), video_id=user_id)
|
||||
|
||||
|
||||
class YoutubeFavouritesIE(InfoExtractor):
|
||||
IE_NAME = 'youtube:favorites'
|
||||
IE_DESC = 'YouTube.com liked videos, ":ytfav" for short (requires authentication)'
|
||||
_VALID_URL = r':ytfav(?:ou?rite)?s?'
|
||||
_LOGIN_REQUIRED = True
|
||||
_TESTS = [{
|
||||
'url': ':ytfav',
|
||||
'only_matching': True,
|
||||
}, {
|
||||
'url': ':ytfavorites',
|
||||
'only_matching': True,
|
||||
}]
|
||||
|
||||
def _real_extract(self, url):
|
||||
return self.url_result(
|
||||
'https://www.youtube.com/playlist?list=LL',
|
||||
ie=YoutubeTabIE.ie_key())
|
||||
|
||||
|
||||
class YoutubeSearchIE(SearchInfoExtractor, YoutubeBaseInfoExtractor):
|
||||
IE_DESC = 'YouTube.com searches'
|
||||
# there doesn't appear to be a real limit, for example if you search for
|
||||
@ -3536,9 +3559,9 @@ class YoutubeFeedsInfoExtractor(YoutubeTabIE):
|
||||
|
||||
class YoutubeWatchLaterIE(InfoExtractor):
|
||||
IE_NAME = 'youtube:watchlater'
|
||||
IE_DESC = 'Youtube watch later list, ":ytwatchlater" or "WL" for short (requires authentication)'
|
||||
_VALID_URL = r'https?://(?:www\.)?youtube\.com/feed/watch_later|:ytwatchlater|WL'
|
||||
|
||||
IE_DESC = 'Youtube watch later list, ":ytwatchlater" for short (requires authentication)'
|
||||
_TESTS = [{
|
||||
'url': 'https://www.youtube.com/feed/watch_later',
|
||||
'only_matching': True,
|
||||
@ -3552,21 +3575,6 @@ class YoutubeWatchLaterIE(InfoExtractor):
|
||||
'https://www.youtube.com/playlist?list=WL', ie=YoutubeTabIE.ie_key())
|
||||
|
||||
|
||||
class YoutubeFavouritesIE(InfoExtractor):
|
||||
IE_NAME = 'youtube:favourites'
|
||||
IE_DESC = 'YouTube.com liked videos, ":ytfav" or "LL" for short (requires authentication)'
|
||||
_VALID_URL = r':ytfav(?:ou?rite)?s?|LL'
|
||||
|
||||
_TESTS = [{
|
||||
'url': ':ytfav',
|
||||
'only_matching': True,
|
||||
}]
|
||||
|
||||
def _real_extract(self, url):
|
||||
return self.url_result(
|
||||
'https://www.youtube.com/playlist?list=LL', ie=YoutubeTabIE.ie_key())
|
||||
|
||||
|
||||
class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
|
||||
IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)'
|
||||
_VALID_URL = r'https?://(?:www\.)?youtube\.com(?:/feed/recommended|/?[?#]|/?$)|:ytrec(?:ommended)?'
|
||||
@ -3575,8 +3583,8 @@ class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
|
||||
|
||||
|
||||
class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
|
||||
IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
|
||||
_VALID_URL = r'https?://(?:www\.)?youtube\.com/feed/subscriptions|:ytsub(?:scription)?s?'
|
||||
IE_DESC = 'YouTube.com subscriptions feed, ":ytsubs" for short (requires authentication)'
|
||||
_FEED_NAME = 'subscriptions'
|
||||
_PLAYLIST_TITLE = 'Youtube Subscriptions'
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user