[yt-dlp.git] / youtube_dl / extractor / soundcloud.py

import json
import re
import itertools

from .common import InfoExtractor
from ..utils import (
    compat_str,
    compat_urlparse,
    compat_urllib_parse,

    ExtractorError,
    unified_strdate,
)


class SoundcloudIE(InfoExtractor):
    """Information extractor for soundcloud.com
       To access the media, the uid of the song and a stream token
       must be extracted from the page source and the script must make
       a request to media.soundcloud.com/crossdomain.xml. Then
       the media can be grabbed by requesting from an url composed
       of the stream token and uid
     """

    _VALID_URL = r'''^(?:https?://)?
                    (?:(?:(?:www\.)?soundcloud\.com/([\w\d-]+)/([\w\d-]+)/?(?:[?].*)?$)
                       |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+))
                       |(?P<widget>w\.soundcloud\.com/player/?.*?url=.*)
                    )
                    '''
    IE_NAME = u'soundcloud'
    _TESTS = [
        {
            u'url': u'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
            u'file': u'62986583.mp3',
            u'md5': u'ebef0a451b909710ed1d7787dddbf0d7',
            u'info_dict': {
                u"upload_date": u"20121011", 
                u"description": u"No Downloads untill we record the finished version this weekend, i was too pumped n i had to post it , earl is prolly gonna b hella p.o'd", 
                u"uploader": u"E.T. ExTerrestrial Music", 
                u"title": u"Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1"
            }
        },
        # not streamable song
        {
            u'url': u'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
            u'info_dict': {
                u'id': u'47127627',
                u'ext': u'mp3',
                u'title': u'Goldrushed',
                u'uploader': u'The Royal Concept',
                u'upload_date': u'20120521',
            },
            u'params': {
                # rtmp
                u'skip_download': True,
            },
        },
    ]

    _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
    _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'

    @classmethod
    def suitable(cls, url):
        return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None

    def report_resolve(self, video_id):
        """Report information extraction."""
        self.to_screen(u'%s: Resolving id' % video_id)

    @classmethod
    def _resolv_url(cls, url):
        return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID

    def _extract_info_dict(self, info, full_title=None, quiet=False):
        track_id = compat_str(info['id'])
        name = full_title or track_id
        if quiet:
            self.report_extraction(name)

        thumbnail = info['artwork_url']
        if thumbnail is not None:
            thumbnail = thumbnail.replace('-large', '-t500x500')
        ext = info.get('original_format', u'mp3')
        result = {
            'id': track_id,
            'uploader': info['user']['username'],
            'upload_date': unified_strdate(info['created_at']),
            'title': info['title'],
            'description': info['description'],
            'thumbnail': thumbnail,
        }
        if info.get('downloadable', False):
            # We can build a direct link to the song
            format_url = (
                u'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
                    track_id, self._CLIENT_ID))
            result['formats'] = [{
                'format_id': 'download',
                'ext': ext,
                'url': format_url,
                'vcodec': 'none',
            }]
        else:
            # We have to retrieve the url
            stream_json = self._download_webpage(
                'http://api.soundcloud.com/i1/tracks/{0}/streams?client_id={1}'.format(track_id, self._IPHONE_CLIENT_ID),
                track_id, u'Downloading track url')

            formats = []
            format_dict = json.loads(stream_json)
            for key, stream_url in format_dict.items():
                if key.startswith(u'http'):
                    formats.append({
                        'format_id': key,
                        'ext': ext,
                        'url': stream_url,
                        'vcodec': 'none',
                    })
                elif key.startswith(u'rtmp'):
                    # The url doesn't have an rtmp app, we have to extract the playpath
                    url, path = stream_url.split('mp3:', 1)
                    formats.append({
                        'format_id': key,
                        'url': url,
                        'play_path': 'mp3:' + path,
                        'ext': ext,
                        'vcodec': 'none',
                    })

            if not formats:
                # We fallback to the stream_url in the original info, this
                # cannot be always used, sometimes it can give an HTTP 404 error
                formats.append({
                    'format_id': u'fallback',
                    'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
                    'ext': ext,
                    'vcodec': 'none',
                })

            def format_pref(f):
                if f['format_id'].startswith('http'):
                    return 2
                if f['format_id'].startswith('rtmp'):
                    return 1
                return 0

            formats.sort(key=format_pref)
            result['formats'] = formats

        return result

    def _real_extract(self, url):
        mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
        if mobj is None:
            raise ExtractorError(u'Invalid URL: %s' % url)

        track_id = mobj.group('track_id')
        if track_id is not None:
            info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
            full_title = track_id
        elif mobj.group('widget'):
            query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
            return self.url_result(query['url'][0], ie='Soundcloud')
        else:
            # extract uploader (which is in the url)
            uploader = mobj.group(1)
            # extract simple title (uploader + slug of song title)
            slug_title =  mobj.group(2)
            full_title = '%s/%s' % (uploader, slug_title)
    
            self.report_resolve(full_title)
    
            url = 'http://soundcloud.com/%s/%s' % (uploader, slug_title)
            info_json_url = self._resolv_url(url)
        info_json = self._download_webpage(info_json_url, full_title, u'Downloading info JSON')

        info = json.loads(info_json)
        return self._extract_info_dict(info, full_title)

class SoundcloudSetIE(SoundcloudIE):
    _VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
    IE_NAME = u'soundcloud:set'
    # it's in tests/test_playlists.py
    _TESTS = []

    def _real_extract(self, url):
        mobj = re.match(self._VALID_URL, url)
        if mobj is None:
            raise ExtractorError(u'Invalid URL: %s' % url)

        # extract uploader (which is in the url)
        uploader = mobj.group(1)
        # extract simple title (uploader + slug of song title)
        slug_title =  mobj.group(2)
        full_title = '%s/sets/%s' % (uploader, slug_title)

        self.report_resolve(full_title)

        url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
        resolv_url = self._resolv_url(url)
        info_json = self._download_webpage(resolv_url, full_title)

        info = json.loads(info_json)
        if 'errors' in info:
            for err in info['errors']:
                self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
            return

        self.report_extraction(full_title)
        return {'_type': 'playlist',
                'entries': [self._extract_info_dict(track) for track in info['tracks']],
                'id': info['id'],
                'title': info['title'],
                }


class SoundcloudUserIE(SoundcloudIE):
    _VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
    IE_NAME = u'soundcloud:user'

    # it's in tests/test_playlists.py
    _TESTS = []

    def _real_extract(self, url):
        mobj = re.match(self._VALID_URL, url)
        uploader = mobj.group('user')

        url = 'http://soundcloud.com/%s/' % uploader
        resolv_url = self._resolv_url(url)
        user_json = self._download_webpage(resolv_url, uploader,
            u'Downloading user info')
        user = json.loads(user_json)

        tracks = []
        for i in itertools.count():
            data = compat_urllib_parse.urlencode({'offset': i*50,
                                                  'client_id': self._CLIENT_ID,
                                                  })
            tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
            response = self._download_webpage(tracks_url, uploader, 
                u'Downloading tracks page %s' % (i+1))
            new_tracks = json.loads(response)
            tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
            if len(new_tracks) < 50:
                break

        return {
            '_type': 'playlist',
            'id': compat_str(user['id']),
            'title': user['username'],
            'entries': tracks,
        }
Commit	Line	Data
aad0d6d5 PH	1	import json
aad0d6d5 PH	2	import re
92790f4e	3	import itertools
aad0d6d5 PH	4
	5	from .common import InfoExtractor
	6	from ..utils import (
	7	compat_str,
668de34c	8	compat_urlparse,
92790f4e	9	compat_urllib_parse,
aad0d6d5 PH	10
	11	ExtractorError,
	12	unified_strdate,
	13	)
	14
	15
	16	class SoundcloudIE(InfoExtractor):
	17	"""Information extractor for soundcloud.com
	18	To access the media, the uid of the song and a stream token
	19	must be extracted from the page source and the script must make
	20	a request to media.soundcloud.com/crossdomain.xml. Then
	21	the media can be grabbed by requesting from an url composed
	22	of the stream token and uid
	23	"""
	24
eb6a41ba JMF	25	_VALID_URL = r'''^(?:https?://)?
	26	(?:(?:(?:www\.)?soundcloud\.com/([\w\d-]+)/([\w\d-]+)/?(?:[?].*)?$)
	27	\|(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+))
c0ade33e	28	\|(?P<widget>w\.soundcloud\.com/player/?.?url=.)
eb6a41ba JMF	29	)
eb6a41ba JMF	30	'''
aad0d6d5	31	IE_NAME = u'soundcloud'
12c167c8 JMF	32	_TESTS = [
	33	{
	34	u'url': u'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
	35	u'file': u'62986583.mp3',
	36	u'md5': u'ebef0a451b909710ed1d7787dddbf0d7',
	37	u'info_dict': {
	38	u"upload_date": u"20121011",
	39	u"description": u"No Downloads untill we record the finished version this weekend, i was too pumped n i had to post it , earl is prolly gonna b hella p.o'd",
	40	u"uploader": u"E.T. ExTerrestrial Music",
	41	u"title": u"Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1"
	42	}
	43	},
	44	# not streamable song
	45	{
	46	u'url': u'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
	47	u'info_dict': {
	48	u'id': u'47127627',
	49	u'ext': u'mp3',
	50	u'title': u'Goldrushed',
	51	u'uploader': u'The Royal Concept',
	52	u'upload_date': u'20120521',
	53	},
	54	u'params': {
	55	# rtmp
	56	u'skip_download': True,
	57	},
	58	},
	59	]
aad0d6d5	60
7d239269	61	_CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
64bb5187	62	_IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
7d239269	63
eb6a41ba JMF	64	@classmethod
	65	def suitable(cls, url):
	66	return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
	67
aad0d6d5 PH	68	def report_resolve(self, video_id):
	69	"""Report information extraction."""
	70	self.to_screen(u'%s: Resolving id' % video_id)
	71
7d239269 JMF	72	@classmethod
	73	def _resolv_url(cls, url):
	74	return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
	75
92790f4e	76	def _extract_info_dict(self, info, full_title=None, quiet=False):
12c167c8 JMF	77	track_id = compat_str(info['id'])
12c167c8 JMF	78	name = full_title or track_id
2a15e706	79	if quiet:
92790f4e	80	self.report_extraction(name)
7d239269 JMF	81
	82	thumbnail = info['artwork_url']
	83	if thumbnail is not None:
	84	thumbnail = thumbnail.replace('-large', '-t500x500')
2a15e706	85	ext = info.get('original_format', u'mp3')
12c167c8	86	result = {
2a15e706	87	'id': track_id,
7d239269 JMF	88	'uploader': info['user']['username'],
7d239269 JMF	89	'upload_date': unified_strdate(info['created_at']),
2a15e706	90	'title': info['title'],
7d239269 JMF	91	'description': info['description'],
	92	'thumbnail': thumbnail,
	93	}
12c167c8	94	if info.get('downloadable', False):
64bb5187	95	# We can build a direct link to the song
2a15e706 PH	96	format_url = (
	97	u'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
	98	track_id, self._CLIENT_ID))
	99	result['formats'] = [{
	100	'format_id': 'download',
	101	'ext': ext,
	102	'url': format_url,
fb04e403	103	'vcodec': 'none',
2a15e706	104	}]
64bb5187 JMF	105	else:
64bb5187 JMF	106	# We have to retrieve the url
12c167c8	107	stream_json = self._download_webpage(
64bb5187	108	'http://api.soundcloud.com/i1/tracks/{0}/streams?client_id={1}'.format(track_id, self._IPHONE_CLIENT_ID),
12c167c8	109	track_id, u'Downloading track url')
2a15e706 PH	110
	111	formats = []
	112	format_dict = json.loads(stream_json)
	113	for key, stream_url in format_dict.items():
	114	if key.startswith(u'http'):
	115	formats.append({
	116	'format_id': key,
	117	'ext': ext,
	118	'url': stream_url,
fb04e403	119	'vcodec': 'none',
2a15e706 PH	120	})
	121	elif key.startswith(u'rtmp'):
	122	# The url doesn't have an rtmp app, we have to extract the playpath
	123	url, path = stream_url.split('mp3:', 1)
	124	formats.append({
	125	'format_id': key,
	126	'url': url,
	127	'play_path': 'mp3:' + path,
	128	'ext': ext,
fb04e403	129	'vcodec': 'none',
2a15e706 PH	130	})
	131
	132	if not formats:
64bb5187 JMF	133	# We fallback to the stream_url in the original info, this
64bb5187 JMF	134	# cannot be always used, sometimes it can give an HTTP 404 error
2a15e706 PH	135	formats.append({
	136	'format_id': u'fallback',
	137	'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
	138	'ext': ext,
fb04e403	139	'vcodec': 'none',
2a15e706 PH	140	})
	141
	142	def format_pref(f):
	143	if f['format_id'].startswith('http'):
	144	return 2
	145	if f['format_id'].startswith('rtmp'):
	146	return 1
	147	return 0
	148
	149	formats.sort(key=format_pref)
	150	result['formats'] = formats
64bb5187	151
12c167c8	152	return result
7d239269	153
aad0d6d5	154	def _real_extract(self, url):
eb6a41ba	155	mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
aad0d6d5 PH	156	if mobj is None:
	157	raise ExtractorError(u'Invalid URL: %s' % url)
	158
eb6a41ba JMF	159	track_id = mobj.group('track_id')
	160	if track_id is not None:
	161	info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
	162	full_title = track_id
668de34c JMF	163	elif mobj.group('widget'):
	164	query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
	165	return self.url_result(query['url'][0], ie='Soundcloud')
eb6a41ba JMF	166	else:
	167	# extract uploader (which is in the url)
	168	uploader = mobj.group(1)
	169	# extract simple title (uploader + slug of song title)
	170	slug_title = mobj.group(2)
	171	full_title = '%s/%s' % (uploader, slug_title)
	172
	173	self.report_resolve(full_title)
	174
	175	url = 'http://soundcloud.com/%s/%s' % (uploader, slug_title)
	176	info_json_url = self._resolv_url(url)
	177	info_json = self._download_webpage(info_json_url, full_title, u'Downloading info JSON')
aad0d6d5 PH	178
aad0d6d5 PH	179	info = json.loads(info_json)
7d239269	180	return self._extract_info_dict(info, full_title)
aad0d6d5	181
7d239269	182	class SoundcloudSetIE(SoundcloudIE):
20db33e2	183	_VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
aad0d6d5	184	IE_NAME = u'soundcloud:set'
12c167c8 JMF	185	# it's in tests/test_playlists.py
12c167c8 JMF	186	_TESTS = []
aad0d6d5	187
aad0d6d5 PH	188	def _real_extract(self, url):
	189	mobj = re.match(self._VALID_URL, url)
	190	if mobj is None:
	191	raise ExtractorError(u'Invalid URL: %s' % url)
	192
	193	# extract uploader (which is in the url)
	194	uploader = mobj.group(1)
	195	# extract simple title (uploader + slug of song title)
	196	slug_title = mobj.group(2)
	197	full_title = '%s/sets/%s' % (uploader, slug_title)
	198
	199	self.report_resolve(full_title)
	200
	201	url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
7d239269	202	resolv_url = self._resolv_url(url)
aad0d6d5 PH	203	info_json = self._download_webpage(resolv_url, full_title)
aad0d6d5 PH	204
aad0d6d5 PH	205	info = json.loads(info_json)
	206	if 'errors' in info:
	207	for err in info['errors']:
	208	self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
	209	return
	210
	211	self.report_extraction(full_title)
7d239269 JMF	212	return {'_type': 'playlist',
	213	'entries': [self._extract_info_dict(track) for track in info['tracks']],
	214	'id': info['id'],
	215	'title': info['title'],
	216	}
92790f4e JMF	217
	218
	219	class SoundcloudUserIE(SoundcloudIE):
c0ade33e	220	_VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
92790f4e JMF	221	IE_NAME = u'soundcloud:user'
	222
	223	# it's in tests/test_playlists.py
12c167c8	224	_TESTS = []
92790f4e JMF	225
	226	def _real_extract(self, url):
	227	mobj = re.match(self._VALID_URL, url)
	228	uploader = mobj.group('user')
	229
	230	url = 'http://soundcloud.com/%s/' % uploader
	231	resolv_url = self._resolv_url(url)
	232	user_json = self._download_webpage(resolv_url, uploader,
	233	u'Downloading user info')
	234	user = json.loads(user_json)
	235
	236	tracks = []
	237	for i in itertools.count():
	238	data = compat_urllib_parse.urlencode({'offset': i*50,
	239	'client_id': self._CLIENT_ID,
	240	})
	241	tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
	242	response = self._download_webpage(tracks_url, uploader,
	243	u'Downloading tracks page %s' % (i+1))
	244	new_tracks = json.loads(response)
	245	tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
	246	if len(new_tracks) < 50:
	247	break
	248
	249	return {
	250	'_type': 'playlist',
	251	'id': compat_str(user['id']),
	252	'title': user['username'],
	253	'entries': tracks,
	254	}