X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=3448bec4fdc96b361c0b071ab12832583533c14f;hb=f1e66cb2eb40b48c6508acbe57207a2d99792bf0;hp=9096a29756ca6e1a66ecd442a92977fa1b999b31;hpb=25f14e9f93295a787e0cb436a5f6179d6174733d;p=youtube-dl
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 9096a2975..3448bec4f 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -785,7 +785,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
s = mobj.group(1)
dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
return '/signature/%s' % dec_s
- dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url)
+ dash_manifest_url = re.sub(r'/s/([a-fA-F0-9\.]+)', decrypt_sig, dash_manifest_url)
dash_doc = self._download_xml(
dash_manifest_url, video_id,
note='Downloading DASH manifest',
@@ -1126,12 +1126,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
self.report_warning(
'Skipping DASH manifest: %r' % e, video_id)
else:
- # Hide the formats we found through non-DASH
+ # Remove the formats we found through non-DASH, they
+ # contain less info and it can be wrong, because we use
+ # fixed values (for example the resolution). See
+ # https://github.com/rg3/youtube-dl/issues/5774 for an
+ # example.
dash_keys = set(df['format_id'] for df in dash_formats)
- for f in formats:
- if f['format_id'] in dash_keys:
- f['format_id'] = 'nondash-%s' % f['format_id']
- f['preference'] = f.get('preference', 0) - 10000
+ formats = [f for f in formats if f['format_id'] not in dash_keys]
formats.extend(dash_formats)
# Check for malformed aspect ratio
@@ -1289,7 +1290,6 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
def _extract_playlist(self, playlist_id):
url = self._TEMPLATE_URL % playlist_id
page = self._download_webpage(url, playlist_id)
- more_widget_html = content_html = page
for match in re.findall(r'([^<]+)
', page):
match = match.strip()
@@ -1309,36 +1309,36 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
self.report_warning('Youtube gives an alert message: ' + match)
# Extract the video ids from the playlist pages
- ids = []
-
- for page_num in itertools.count(1):
- matches = re.finditer(self._VIDEO_RE, content_html)
- # We remove the duplicates and the link with index 0
- # (it's not the first video of the playlist)
- new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
- ids.extend(new_ids)
-
- mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html)
- if not mobj:
- break
+ def _entries():
+ more_widget_html = content_html = page
+ for page_num in itertools.count(1):
+ matches = re.finditer(self._VIDEO_RE, content_html)
+ # We remove the duplicates and the link with index 0
+ # (it's not the first video of the playlist)
+ new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
+ for vid_id in new_ids:
+ yield self.url_result(vid_id, 'Youtube', video_id=vid_id)
+
+ mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html)
+ if not mobj:
+ break
- more = self._download_json(
- 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
- 'Downloading page #%s' % page_num,
- transform_source=uppercase_escape)
- content_html = more['content_html']
- if not content_html.strip():
- # Some webpages show a "Load more" button but they don't
- # have more videos
- break
- more_widget_html = more['load_more_widget_html']
+ more = self._download_json(
+ 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
+ 'Downloading page #%s' % page_num,
+ transform_source=uppercase_escape)
+ content_html = more['content_html']
+ if not content_html.strip():
+ # Some webpages show a "Load more" button but they don't
+ # have more videos
+ break
+ more_widget_html = more['load_more_widget_html']
playlist_title = self._html_search_regex(
r'(?s)',
page, 'title')
- url_results = self._ids_to_results(ids)
- return self.playlist_result(url_results, playlist_id, playlist_title)
+ return self.playlist_result(_entries(), playlist_id, playlist_title)
def _real_extract(self, url):
# Extract playlist id
@@ -1398,6 +1398,24 @@ class YoutubeChannelIE(InfoExtractor):
channel_id = self._match_id(url)
url = self._TEMPLATE_URL % channel_id
+
+ # Channel by page listing is restricted to 35 pages of 30 items, i.e. 1050 videos total (see #5778)
+ # Workaround by extracting as a playlist if managed to obtain channel playlist URL
+ # otherwise fallback on channel by page extraction
+ channel_page = self._download_webpage(
+ url + '?view=57', channel_id,
+ 'Downloading channel page', fatal=False)
+ channel_playlist_id = self._html_search_meta(
+ 'channelId', channel_page, 'channel id', default=None)
+ if not channel_playlist_id:
+ channel_playlist_id = self._search_regex(
+ r'data-channel-external-id="([^"]+)"',
+ channel_page, 'channel id', default=None)
+ if channel_playlist_id and channel_playlist_id.startswith('UC'):
+ playlist_id = 'UU' + channel_playlist_id[2:]
+ return self.url_result(
+ compat_urlparse.urljoin(url, '/playlist?list=%s' % playlist_id), 'YoutubePlaylist')
+
channel_page = self._download_webpage(url, channel_id, 'Downloading page #1')
autogenerated = re.search(r'''(?x)
class="[^"]*?(?:
@@ -1621,10 +1639,16 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
# for the video ids doesn't contain an index
ids = []
more_widget_html = content_html = page
-
for page_num in itertools.count(1):
matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
- new_ids = orderedSet(matches)
+
+ # 'recommended' feed has infinite 'load more' and each new portion spins
+ # the same videos in (sometimes) slightly different order, so we'll check
+ # for unicity and break when portion has no new videos
+ new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches))
+ if not new_ids:
+ break
+
ids.extend(new_ids)
mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html)