From 2aae2c91ff0608d02c43058485224b10d30b892d Mon Sep 17 00:00:00 2001 From: Abdullah Ibn Fulan <54185653+abdullah-if@users.noreply.github.com> Date: Mon, 20 Dec 2021 07:53:42 +0600 Subject: [PATCH] [audiomack] Update album and song VALID_URL (#1203) Original PR: https://github.com/ytdl-org/youtube-dl/pull/29810 Closes #1352, https://github.com/ytdl-org/youtube-dl/issues/29800 Authored by: abdullah-if, dirkf --- yt_dlp/extractor/audiomack.py | 34 +++++++++++++++++++--------------- 1 file changed, 19 insertions(+), 15 deletions(-) diff --git a/yt_dlp/extractor/audiomack.py b/yt_dlp/extractor/audiomack.py index cc7771354c..31fb859ae3 100644 --- a/yt_dlp/extractor/audiomack.py +++ b/yt_dlp/extractor/audiomack.py @@ -14,7 +14,7 @@ from ..utils import ( class AudiomackIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?audiomack\.com/song/(?P[\w/-]+)' + _VALID_URL = r'https?://(?:www\.)?audiomack\.com/(?:song/|(?=.+/song/))(?P[\w/-]+)' IE_NAME = 'audiomack' _TESTS = [ # hosted on audiomack @@ -39,15 +39,16 @@ class AudiomackIE(InfoExtractor): 'title': 'Black Mamba Freestyle [Prod. By Danny Wolf]', 'uploader': 'ILOVEMAKONNEN', 'upload_date': '20160414', - } + }, + 'skip': 'Song has been removed from the site', }, ] def _real_extract(self, url): - # URLs end with [uploader name]/[uploader title] + # URLs end with [uploader name]/song/[uploader title] # this title is whatever the user types in, and is rarely # the proper song title. Real metadata is in the api response - album_url_tag = self._match_id(url) + album_url_tag = self._match_id(url).replace('/song/', '/') # Request the extended version of the api for extra fields like artist and title api_response = self._download_json( @@ -73,13 +74,13 @@ class AudiomackIE(InfoExtractor): class AudiomackAlbumIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?audiomack\.com/album/(?P[\w/-]+)' + _VALID_URL = r'https?://(?:www\.)?audiomack\.com/(?:album/|(?=.+/album/))(?P[\w/-]+)' IE_NAME = 'audiomack:album' _TESTS = [ # Standard album playlist { 'url': 'http://www.audiomack.com/album/flytunezcom/tha-tour-part-2-mixtape', - 'playlist_count': 15, + 'playlist_count': 11, 'info_dict': { 'id': '812251', @@ -95,24 +96,27 @@ class AudiomackAlbumIE(InfoExtractor): }, 'playlist': [{ 'info_dict': { - 'title': 'PPP (Pistol P Project) - 9. Heaven or Hell (CHIMACA) ft Zuse (prod by DJ FU)', - 'id': '837577', + 'title': 'PPP (Pistol P Project) - 8. Real (prod by SYK SENSE )', + 'id': '837576', + 'ext': 'mp3', + 'uploader': 'Lil Herb a.k.a. G Herbo', + } + }, { + 'info_dict': { + 'title': 'PPP (Pistol P Project) - 10. 4 Minutes Of Hell Part 4 (prod by DY OF 808 MAFIA)', + 'id': '837580', 'ext': 'mp3', 'uploader': 'Lil Herb a.k.a. G Herbo', } }], - 'params': { - 'playliststart': 9, - 'playlistend': 9, - } } ] def _real_extract(self, url): - # URLs end with [uploader name]/[uploader title] + # URLs end with [uploader name]/album/[uploader title] # this title is whatever the user types in, and is rarely # the proper song title. Real metadata is in the api response - album_url_tag = self._match_id(url) + album_url_tag = self._match_id(url).replace('/album/', '/') result = {'_type': 'playlist', 'entries': []} # There is no one endpoint for album metadata - instead it is included/repeated in each song's metadata # Therefore we don't know how many songs the album has and must infi-loop until failure @@ -134,7 +138,7 @@ class AudiomackAlbumIE(InfoExtractor): # Pull out the album metadata and add to result (if it exists) for resultkey, apikey in [('id', 'album_id'), ('title', 'album_title')]: if apikey in api_response and resultkey not in result: - result[resultkey] = api_response[apikey] + result[resultkey] = compat_str(api_response[apikey]) song_id = url_basename(api_response['url']).rpartition('.')[0] result['entries'].append({ 'id': compat_str(api_response.get('id', song_id)),