mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-30 07:28:19 +01:00
197d0b03b6
Closes #11361 Authored by: avagordon01, bashonly, grqz, Grub4K, seproDev Co-authored-by: Ava Gordon <avagordon01@gmail.com> Co-authored-by: bashonly <bashonly@protonmail.com> Co-authored-by: N/Ame <173015200+grqz@users.noreply.github.com> Co-authored-by: Simon Sawicki <contact@grub4k.xyz>
522 lines
20 KiB
Python
522 lines
20 KiB
Python
import json
|
|
import random
|
|
import re
|
|
import time
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
KNOWN_EXTENSIONS,
|
|
ExtractorError,
|
|
extract_attributes,
|
|
float_or_none,
|
|
int_or_none,
|
|
parse_filesize,
|
|
str_or_none,
|
|
try_get,
|
|
unified_strdate,
|
|
unified_timestamp,
|
|
update_url_query,
|
|
url_or_none,
|
|
urljoin,
|
|
)
|
|
from ..utils.traversal import find_element, traverse_obj
|
|
|
|
|
|
class BandcampIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?P<uploader>[^/]+)\.bandcamp\.com/track/(?P<id>[^/?#&]+)'
|
|
_EMBED_REGEX = [r'<meta property="og:url"[^>]*?content="(?P<url>.*?bandcamp\.com.*?)"']
|
|
_TESTS = [{
|
|
'url': 'http://youtube-dl.bandcamp.com/track/youtube-dl-test-song',
|
|
'md5': 'c557841d5e50261777a6585648adf439',
|
|
'info_dict': {
|
|
'id': '1812978515',
|
|
'ext': 'mp3',
|
|
'title': 'youtube-dl "\'/\\ä↭ - youtube-dl "\'/\\ä↭ - youtube-dl test song "\'/\\ä↭',
|
|
'duration': 9.8485,
|
|
'uploader': 'youtube-dl "\'/\\ä↭',
|
|
'upload_date': '20121129',
|
|
'timestamp': 1354224127,
|
|
'track': 'youtube-dl "\'/\\ä↭ - youtube-dl test song "\'/\\ä↭',
|
|
'album_artist': 'youtube-dl "\'/\\ä↭',
|
|
'track_id': '1812978515',
|
|
'artist': 'youtube-dl "\'/\\ä↭',
|
|
'uploader_url': 'https://youtube-dl.bandcamp.com',
|
|
'uploader_id': 'youtube-dl',
|
|
'thumbnail': 'https://f4.bcbits.com/img/a3216802731_5.jpg',
|
|
'artists': ['youtube-dl "\'/\\ä↭'],
|
|
'album_artists': ['youtube-dl "\'/\\ä↭'],
|
|
},
|
|
'skip': 'There is a limit of 200 free downloads / month for the test song',
|
|
}, {
|
|
# free download
|
|
'url': 'http://benprunty.bandcamp.com/track/lanius-battle',
|
|
'info_dict': {
|
|
'id': '2650410135',
|
|
'ext': 'm4a',
|
|
'acodec': r're:[fa]lac',
|
|
'title': 'Ben Prunty - Lanius (Battle)',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'uploader': 'Ben Prunty',
|
|
'timestamp': 1396508491,
|
|
'upload_date': '20140403',
|
|
'release_timestamp': 1396483200,
|
|
'release_date': '20140403',
|
|
'duration': 260.877,
|
|
'track': 'Lanius (Battle)',
|
|
'track_number': 1,
|
|
'track_id': '2650410135',
|
|
'artist': 'Ben Prunty',
|
|
'album_artist': 'Ben Prunty',
|
|
'album': 'FTL: Advanced Edition Soundtrack',
|
|
'uploader_url': 'https://benprunty.bandcamp.com',
|
|
'uploader_id': 'benprunty',
|
|
},
|
|
}, {
|
|
# no free download, mp3 128
|
|
'url': 'https://relapsealumni.bandcamp.com/track/hail-to-fire',
|
|
'md5': 'fec12ff55e804bb7f7ebeb77a800c8b7',
|
|
'info_dict': {
|
|
'id': '2584466013',
|
|
'ext': 'mp3',
|
|
'title': 'Mastodon - Hail to Fire',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'uploader': 'Mastodon',
|
|
'timestamp': 1322005399,
|
|
'upload_date': '20111122',
|
|
'release_timestamp': 1076112000,
|
|
'release_date': '20040207',
|
|
'duration': 120.79,
|
|
'track': 'Hail to Fire',
|
|
'track_number': 5,
|
|
'track_id': '2584466013',
|
|
'artist': 'Mastodon',
|
|
'album_artist': 'Mastodon',
|
|
'album': 'Call of the Mastodon',
|
|
'uploader_url': 'https://relapsealumni.bandcamp.com',
|
|
'uploader_id': 'relapsealumni',
|
|
},
|
|
}, {
|
|
# track from compilation album (artist/album_artist difference)
|
|
'url': 'https://diskotopia.bandcamp.com/track/safehouse',
|
|
'md5': '19c5337bca1428afa54129f86a2f6a69',
|
|
'info_dict': {
|
|
'id': '1978174799',
|
|
'ext': 'mp3',
|
|
'title': 'submerse - submerse - Safehouse',
|
|
'thumbnail': r're:^https?://.*\.jpg$',
|
|
'uploader': 'submerse',
|
|
'timestamp': 1480779297,
|
|
'upload_date': '20161203',
|
|
'release_timestamp': 1481068800,
|
|
'release_date': '20161207',
|
|
'duration': 154.066,
|
|
'track': 'submerse - Safehouse',
|
|
'track_number': 3,
|
|
'track_id': '1978174799',
|
|
'artist': 'submerse',
|
|
'album_artist': 'Diskotopia',
|
|
'album': 'DSK F/W 2016-2017 Free Compilation',
|
|
'uploader_url': 'https://diskotopia.bandcamp.com',
|
|
'uploader_id': 'diskotopia',
|
|
},
|
|
}]
|
|
|
|
def _extract_data_attr(self, webpage, video_id, attr='tralbum', fatal=True):
|
|
return self._parse_json(self._html_search_regex(
|
|
rf'data-{attr}=(["\'])({{.+?}})\1', webpage,
|
|
attr + ' data', group=2), video_id, fatal=fatal)
|
|
|
|
def _real_extract(self, url):
|
|
title, uploader = self._match_valid_url(url).group('id', 'uploader')
|
|
webpage = self._download_webpage(url, title)
|
|
tralbum = self._extract_data_attr(webpage, title)
|
|
thumbnail = self._og_search_thumbnail(webpage)
|
|
|
|
track_id = None
|
|
track = None
|
|
track_number = None
|
|
duration = None
|
|
|
|
formats = []
|
|
track_info = try_get(tralbum, lambda x: x['trackinfo'][0], dict)
|
|
if track_info:
|
|
file_ = track_info.get('file')
|
|
if isinstance(file_, dict):
|
|
for format_id, format_url in file_.items():
|
|
if not url_or_none(format_url):
|
|
continue
|
|
ext, abr_str = format_id.split('-', 1)
|
|
formats.append({
|
|
'format_id': format_id,
|
|
'url': self._proto_relative_url(format_url, 'http:'),
|
|
'ext': ext,
|
|
'vcodec': 'none',
|
|
'acodec': ext,
|
|
'abr': int_or_none(abr_str),
|
|
})
|
|
track = track_info.get('title')
|
|
track_id = str_or_none(
|
|
track_info.get('track_id') or track_info.get('id'))
|
|
track_number = int_or_none(track_info.get('track_num'))
|
|
duration = float_or_none(track_info.get('duration'))
|
|
|
|
embed = self._extract_data_attr(webpage, title, 'embed', False)
|
|
current = tralbum.get('current') or {}
|
|
artist = embed.get('artist') or current.get('artist') or tralbum.get('artist')
|
|
album_artist = self._html_search_regex(
|
|
r'<h3 class="albumTitle">[\S\s]*?by\s*<span>\s*<a href="[^>]+">\s*([^>]+?)\s*</a>',
|
|
webpage, 'album artist', fatal=False)
|
|
timestamp = unified_timestamp(
|
|
current.get('publish_date') or tralbum.get('album_publish_date'))
|
|
|
|
download_link = tralbum.get('freeDownloadPage')
|
|
if download_link:
|
|
track_id = str(tralbum['id'])
|
|
|
|
download_webpage = self._download_webpage(
|
|
download_link, track_id, 'Downloading free downloads page')
|
|
|
|
blob = self._extract_data_attr(download_webpage, track_id, 'blob')
|
|
|
|
info = try_get(
|
|
blob, (lambda x: x['digital_items'][0],
|
|
lambda x: x['download_items'][0]), dict)
|
|
if info:
|
|
downloads = info.get('downloads')
|
|
if isinstance(downloads, dict):
|
|
if not track:
|
|
track = info.get('title')
|
|
if not artist:
|
|
artist = info.get('artist')
|
|
if not thumbnail:
|
|
thumbnail = info.get('thumb_url')
|
|
|
|
download_formats = {}
|
|
download_formats_list = blob.get('download_formats')
|
|
if isinstance(download_formats_list, list):
|
|
for f in blob['download_formats']:
|
|
name, ext = f.get('name'), f.get('file_extension')
|
|
if all(isinstance(x, str) for x in (name, ext)):
|
|
download_formats[name] = ext.strip('.')
|
|
|
|
for format_id, f in downloads.items():
|
|
format_url = f.get('url')
|
|
if not format_url:
|
|
continue
|
|
# Stat URL generation algorithm is reverse engineered from
|
|
# download_*_bundle_*.js
|
|
stat_url = update_url_query(
|
|
format_url.replace('/download/', '/statdownload/'), {
|
|
'.rand': int(time.time() * 1000 * random.random()),
|
|
})
|
|
format_id = f.get('encoding_name') or format_id
|
|
stat = self._download_json(
|
|
stat_url, track_id, f'Downloading {format_id} JSON',
|
|
transform_source=lambda s: s[s.index('{'):s.rindex('}') + 1],
|
|
fatal=False)
|
|
if not stat:
|
|
continue
|
|
retry_url = url_or_none(stat.get('retry_url'))
|
|
if not retry_url:
|
|
continue
|
|
formats.append({
|
|
'url': self._proto_relative_url(retry_url, 'http:'),
|
|
'ext': download_formats.get(format_id),
|
|
'format_id': format_id,
|
|
'format_note': f.get('description'),
|
|
'filesize': parse_filesize(f.get('size_mb')),
|
|
'vcodec': 'none',
|
|
'acodec': format_id.split('-')[0],
|
|
})
|
|
|
|
title = f'{artist} - {track}' if artist else track
|
|
|
|
if not duration:
|
|
duration = float_or_none(self._html_search_meta(
|
|
'duration', webpage, default=None))
|
|
|
|
return {
|
|
'id': track_id,
|
|
'title': title,
|
|
'thumbnail': thumbnail,
|
|
'uploader': artist,
|
|
'uploader_id': uploader,
|
|
'uploader_url': f'https://{uploader}.bandcamp.com',
|
|
'timestamp': timestamp,
|
|
'release_timestamp': unified_timestamp(tralbum.get('album_release_date')),
|
|
'duration': duration,
|
|
'track': track,
|
|
'track_number': track_number,
|
|
'track_id': track_id,
|
|
'artist': artist,
|
|
'album': embed.get('album_title'),
|
|
'album_artist': album_artist,
|
|
'formats': formats,
|
|
}
|
|
|
|
|
|
class BandcampAlbumIE(BandcampIE): # XXX: Do not subclass from concrete IE
|
|
IE_NAME = 'Bandcamp:album'
|
|
_VALID_URL = r'https?://(?:(?P<subdomain>[^.]+)\.)?bandcamp\.com/album/(?P<id>[^/?#&]+)'
|
|
|
|
_TESTS = [{
|
|
'url': 'http://blazo.bandcamp.com/album/jazz-format-mixtape-vol-1',
|
|
'playlist': [
|
|
{
|
|
'md5': '39bc1eded3476e927c724321ddf116cf',
|
|
'info_dict': {
|
|
'id': '1353101989',
|
|
'ext': 'mp3',
|
|
'title': 'Blazo - Intro',
|
|
'timestamp': 1311756226,
|
|
'upload_date': '20110727',
|
|
'uploader': 'Blazo',
|
|
'thumbnail': 'https://f4.bcbits.com/img/a1721150828_5.jpg',
|
|
'album_artists': ['Blazo'],
|
|
'uploader_url': 'https://blazo.bandcamp.com',
|
|
'release_date': '20110727',
|
|
'release_timestamp': 1311724800.0,
|
|
'track': 'Intro',
|
|
'uploader_id': 'blazo',
|
|
'track_number': 1,
|
|
'album': 'Jazz Format Mixtape vol.1',
|
|
'artists': ['Blazo'],
|
|
'duration': 19.335,
|
|
'track_id': '1353101989',
|
|
},
|
|
},
|
|
{
|
|
'md5': '1a2c32e2691474643e912cc6cd4bffaa',
|
|
'info_dict': {
|
|
'id': '38097443',
|
|
'ext': 'mp3',
|
|
'title': 'Blazo - Kero One - Keep It Alive (Blazo remix)',
|
|
'timestamp': 1311757238,
|
|
'upload_date': '20110727',
|
|
'uploader': 'Blazo',
|
|
'track': 'Kero One - Keep It Alive (Blazo remix)',
|
|
'release_date': '20110727',
|
|
'track_id': '38097443',
|
|
'track_number': 2,
|
|
'duration': 181.467,
|
|
'uploader_url': 'https://blazo.bandcamp.com',
|
|
'album': 'Jazz Format Mixtape vol.1',
|
|
'uploader_id': 'blazo',
|
|
'album_artists': ['Blazo'],
|
|
'artists': ['Blazo'],
|
|
'thumbnail': 'https://f4.bcbits.com/img/a1721150828_5.jpg',
|
|
'release_timestamp': 1311724800.0,
|
|
},
|
|
},
|
|
],
|
|
'info_dict': {
|
|
'title': 'Jazz Format Mixtape vol.1',
|
|
'id': 'jazz-format-mixtape-vol-1',
|
|
'uploader_id': 'blazo',
|
|
'description': 'md5:38052a93217f3ffdc033cd5dbbce2989',
|
|
},
|
|
'params': {
|
|
'playlistend': 2,
|
|
},
|
|
'skip': 'Bandcamp imposes download limits.',
|
|
}, {
|
|
'url': 'http://nightbringer.bandcamp.com/album/hierophany-of-the-open-grave',
|
|
'info_dict': {
|
|
'title': 'Hierophany of the Open Grave',
|
|
'uploader_id': 'nightbringer',
|
|
'id': 'hierophany-of-the-open-grave',
|
|
},
|
|
'playlist_mincount': 9,
|
|
}, {
|
|
# with escaped quote in title
|
|
'url': 'https://jstrecords.bandcamp.com/album/entropy-ep',
|
|
'info_dict': {
|
|
'title': '"Entropy" EP',
|
|
'uploader_id': 'jstrecords',
|
|
'id': 'entropy-ep',
|
|
'description': 'md5:0ff22959c943622972596062f2f366a5',
|
|
},
|
|
'playlist_mincount': 3,
|
|
}, {
|
|
# not all tracks have songs
|
|
'url': 'https://insulters.bandcamp.com/album/we-are-the-plague',
|
|
'info_dict': {
|
|
'id': 'we-are-the-plague',
|
|
'title': 'WE ARE THE PLAGUE',
|
|
'uploader_id': 'insulters',
|
|
'description': 'md5:b3cf845ee41b2b1141dc7bde9237255f',
|
|
},
|
|
'playlist_count': 2,
|
|
}]
|
|
|
|
@classmethod
|
|
def suitable(cls, url):
|
|
return (False
|
|
if BandcampWeeklyIE.suitable(url) or BandcampIE.suitable(url)
|
|
else super().suitable(url))
|
|
|
|
def _real_extract(self, url):
|
|
uploader_id, album_id = self._match_valid_url(url).groups()
|
|
playlist_id = album_id or uploader_id
|
|
webpage = self._download_webpage(url, playlist_id)
|
|
tralbum = self._extract_data_attr(webpage, playlist_id)
|
|
track_info = tralbum.get('trackinfo')
|
|
if not track_info:
|
|
raise ExtractorError('The page doesn\'t contain any tracks')
|
|
# Only tracks with duration info have songs
|
|
entries = [
|
|
self.url_result(
|
|
urljoin(url, t['title_link']), BandcampIE.ie_key(),
|
|
str_or_none(t.get('track_id') or t.get('id')), t.get('title'))
|
|
for t in track_info
|
|
if t.get('duration')]
|
|
|
|
current = tralbum.get('current') or {}
|
|
|
|
return {
|
|
'_type': 'playlist',
|
|
'uploader_id': uploader_id,
|
|
'id': playlist_id,
|
|
'title': current.get('title'),
|
|
'description': current.get('about'),
|
|
'entries': entries,
|
|
}
|
|
|
|
|
|
class BandcampWeeklyIE(BandcampIE): # XXX: Do not subclass from concrete IE
|
|
IE_NAME = 'Bandcamp:weekly'
|
|
_VALID_URL = r'https?://(?:www\.)?bandcamp\.com/?\?(?:.*?&)?show=(?P<id>\d+)'
|
|
_TESTS = [{
|
|
'url': 'https://bandcamp.com/?show=224',
|
|
'md5': '61acc9a002bed93986b91168aa3ab433',
|
|
'info_dict': {
|
|
'id': '224',
|
|
'ext': 'mp3',
|
|
'title': 'BC Weekly April 4th 2017 - Magic Moments',
|
|
'description': 'md5:5d48150916e8e02d030623a48512c874',
|
|
'duration': 5829.77,
|
|
'release_date': '20170404',
|
|
'series': 'Bandcamp Weekly',
|
|
'episode': 'Magic Moments',
|
|
'episode_id': '224',
|
|
},
|
|
'params': {
|
|
'format': 'mp3-128',
|
|
},
|
|
}, {
|
|
'url': 'https://bandcamp.com/?blah/blah@&show=228',
|
|
'only_matching': True,
|
|
}]
|
|
|
|
def _real_extract(self, url):
|
|
show_id = self._match_id(url)
|
|
webpage = self._download_webpage(url, show_id)
|
|
|
|
blob = self._extract_data_attr(webpage, show_id, 'blob')
|
|
|
|
show = blob['bcw_data'][show_id]
|
|
|
|
formats = []
|
|
for format_id, format_url in show['audio_stream'].items():
|
|
if not url_or_none(format_url):
|
|
continue
|
|
for known_ext in KNOWN_EXTENSIONS:
|
|
if known_ext in format_id:
|
|
ext = known_ext
|
|
break
|
|
else:
|
|
ext = None
|
|
formats.append({
|
|
'format_id': format_id,
|
|
'url': format_url,
|
|
'ext': ext,
|
|
'vcodec': 'none',
|
|
})
|
|
|
|
title = show.get('audio_title') or 'Bandcamp Weekly'
|
|
subtitle = show.get('subtitle')
|
|
if subtitle:
|
|
title += f' - {subtitle}'
|
|
|
|
return {
|
|
'id': show_id,
|
|
'title': title,
|
|
'description': show.get('desc') or show.get('short_desc'),
|
|
'duration': float_or_none(show.get('audio_duration')),
|
|
'is_live': False,
|
|
'release_date': unified_strdate(show.get('published_date')),
|
|
'series': 'Bandcamp Weekly',
|
|
'episode': show.get('subtitle'),
|
|
'episode_id': show_id,
|
|
'formats': formats,
|
|
}
|
|
|
|
|
|
class BandcampUserIE(InfoExtractor):
|
|
IE_NAME = 'Bandcamp:user'
|
|
_VALID_URL = r'https?://(?!www\.)(?P<id>[^.]+)\.bandcamp\.com(?:/music)?/?(?:[#?]|$)'
|
|
|
|
_TESTS = [{
|
|
# Type 1 Bandcamp user page.
|
|
'url': 'https://adrianvonziegler.bandcamp.com',
|
|
'info_dict': {
|
|
'id': 'adrianvonziegler',
|
|
'title': 'Discography of adrianvonziegler',
|
|
},
|
|
'playlist_mincount': 23,
|
|
}, {
|
|
# Bandcamp user page with only one album
|
|
'url': 'http://dotscale.bandcamp.com',
|
|
'info_dict': {
|
|
'id': 'dotscale',
|
|
'title': 'Discography of dotscale',
|
|
},
|
|
'playlist_count': 1,
|
|
}, {
|
|
# Type 2 Bandcamp user page.
|
|
'url': 'https://nightcallofficial.bandcamp.com',
|
|
'info_dict': {
|
|
'id': 'nightcallofficial',
|
|
'title': 'Discography of nightcallofficial',
|
|
},
|
|
'playlist_count': 4,
|
|
}, {
|
|
'url': 'https://steviasphere.bandcamp.com/music',
|
|
'playlist_mincount': 47,
|
|
'info_dict': {
|
|
'id': 'steviasphere',
|
|
'title': 'Discography of steviasphere',
|
|
},
|
|
}, {
|
|
'url': 'https://coldworldofficial.bandcamp.com/music',
|
|
'playlist_mincount': 7,
|
|
'info_dict': {
|
|
'id': 'coldworldofficial',
|
|
'title': 'Discography of coldworldofficial',
|
|
},
|
|
}, {
|
|
'url': 'https://nuclearwarnowproductions.bandcamp.com/music',
|
|
'playlist_mincount': 399,
|
|
'info_dict': {
|
|
'id': 'nuclearwarnowproductions',
|
|
'title': 'Discography of nuclearwarnowproductions',
|
|
},
|
|
}]
|
|
|
|
def _yield_items(self, webpage):
|
|
yield from (
|
|
re.findall(r'<li data-item-id=["\'][^>]+>\s*<a href=["\'](?![^"\'/]*?/merch)([^"\']+)', webpage)
|
|
or re.findall(r'<div[^>]+trackTitle["\'][^"\']+["\']([^"\']+)', webpage))
|
|
|
|
yield from traverse_obj(webpage, (
|
|
{find_element(id='music-grid', html=True)}, {extract_attributes},
|
|
'data-client-items', {json.loads}, ..., 'page_url', {str}))
|
|
|
|
def _real_extract(self, url):
|
|
uploader = self._match_id(url)
|
|
webpage = self._download_webpage(url, uploader)
|
|
|
|
return self.playlist_from_matches(
|
|
self._yield_items(webpage), uploader, f'Discography of {uploader}',
|
|
getter=urljoin(url))
|