From d0ba55871e6754fdc8a6a28543581989ba3c50fe Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sat, 25 Mar 2017 01:17:17 +0700 Subject: [PATCH] [youtube] Improve _VALID_URLs (closes #12538) --- youtube_dl/extractor/youtube.py | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index caa048249..ca40de522 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -59,6 +59,8 @@ 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)[0-9A-Za-z-_]{10,}' + def _set_language(self): self._set_cookie( '.youtube.com', 'PREF', 'f1=50000000&hl=en', @@ -265,9 +267,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor): ) )? # all until now is optional -> you can pass the naked ID ([0-9A-Za-z_-]{11}) # here is it! the YouTube video ID - (?!.*?\blist=) # combined list/video URLs are handled by the playlist IE + (?!.*?\blist= + (?: + %(playlist_id)s| # combined list/video URLs are handled by the playlist IE + WL # WL are handled by the watch later IE + ) + ) (?(1).+)? # if we found the ID, everything can follow - $""" + $""" % {'playlist_id': YoutubeBaseInfoExtractor._PLAYLIST_ID_RE} _NEXT_URL_RE = r'[\?&]next_url=([^&]+)' _formats = { '5': {'ext': 'flv', 'width': 400, 'height': 240, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'}, @@ -924,6 +931,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'url': 'sJL6WA-aGkQ', 'only_matching': True, }, + { + 'url': 'https://www.youtube.com/watch?v=MuAGGZNfUkU&list=RDMM', + 'only_matching': True, + }, ] def __init__(self, *args, **kwargs): @@ -1864,8 +1875,8 @@ class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor): ) .* | - ((?:PL|LL|EC|UU|FL|RD|UL|TL)[0-9A-Za-z-_]{10,}) - )""" + (%(playlist_id)s) + )""" % {'playlist_id': YoutubeBaseInfoExtractor._PLAYLIST_ID_RE} _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&disable_polymer=true' _VIDEO_RE = r'href="\s*/watch\?v=(?P[0-9A-Za-z_-]{11})&[^"]*?index=(?P\d+)(?:[^>]+>(?P[^<]+))?' IE_NAME = 'youtube:playlist'