diff options
Diffstat (limited to 'hypervideo_dl/extractor/soundcloud.py')
| -rw-r--r-- | hypervideo_dl/extractor/soundcloud.py | 370 |
1 files changed, 178 insertions, 192 deletions
diff --git a/hypervideo_dl/extractor/soundcloud.py b/hypervideo_dl/extractor/soundcloud.py index 92535f7..c2344dd 100644 --- a/hypervideo_dl/extractor/soundcloud.py +++ b/hypervideo_dl/extractor/soundcloud.py @@ -1,6 +1,3 @@ -# coding: utf-8 -from __future__ import unicode_literals - import itertools import re import json @@ -12,7 +9,6 @@ from .common import ( ) from ..compat import ( compat_HTTPError, - compat_kwargs, compat_str, ) from ..utils import ( @@ -23,7 +19,6 @@ from ..utils import ( int_or_none, KNOWN_EXTENSIONS, mimetype2ext, - remove_end, parse_qs, str_or_none, try_get, @@ -37,18 +32,13 @@ from ..utils import ( class SoundcloudEmbedIE(InfoExtractor): _VALID_URL = r'https?://(?:w|player|p)\.soundcloud\.com/player/?.*?\burl=(?P<id>.+)' + _EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1'] _TEST = { # from https://www.soundi.fi/uutiset/ennakkokuuntelussa-timo-kaukolammen-station-to-station-to-station-julkaisua-juhlitaan-tanaan-g-livelabissa/ 'url': 'https://w.soundcloud.com/player/?visual=true&url=https%3A%2F%2Fapi.soundcloud.com%2Fplaylists%2F922213810&show_artwork=true&maxwidth=640&maxheight=960&dnt=1&secret_token=s-ziYey', 'only_matching': True, } - @staticmethod - def _extract_urls(webpage): - return [m.group('url') for m in re.finditer( - r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1', - webpage)] - def _real_extract(self, url): query = parse_qs(url) api_url = query['url'][0] @@ -70,8 +60,23 @@ class SoundcloudBaseIE(InfoExtractor): _access_token = None _HEADERS = {} + _IMAGE_REPL_RE = r'-([0-9a-z]+)\.jpg' + + _ARTWORK_MAP = { + 'mini': 16, + 'tiny': 20, + 'small': 32, + 'badge': 47, + 't67x67': 67, + 'large': 100, + 't300x300': 300, + 'crop': 400, + 't500x500': 500, + 'original': 0, + } + def _store_client_id(self, client_id): - self._downloader.cache.store('soundcloud', 'client_id', client_id) + self.cache.store('soundcloud', 'client_id', client_id) def _update_client_id(self): webpage = self._download_webpage('https://soundcloud.com/', None) @@ -96,7 +101,7 @@ class SoundcloudBaseIE(InfoExtractor): query['client_id'] = self._CLIENT_ID kwargs['query'] = query try: - return super()._download_json(*args, **compat_kwargs(kwargs)) + return super()._download_json(*args, **kwargs) except ExtractorError as e: if isinstance(e.cause, compat_HTTPError) and e.cause.code in (401, 403): self._store_client_id(None) @@ -108,7 +113,7 @@ class SoundcloudBaseIE(InfoExtractor): raise def _initialize_pre_login(self): - self._CLIENT_ID = self._downloader.cache.load('soundcloud', 'client_id') or 'a3e059563d7fd3372b49b37f00a00bcf' + self._CLIENT_ID = self.cache.load('soundcloud', 'client_id') or 'a3e059563d7fd3372b49b37f00a00bcf' def _perform_login(self, username, password): if username != 'oauth': @@ -189,6 +194,157 @@ class SoundcloudBaseIE(InfoExtractor): return out + def _extract_info_dict(self, info, full_title=None, secret_token=None, extract_flat=False): + track_id = compat_str(info['id']) + title = info['title'] + + format_urls = set() + formats = [] + query = {'client_id': self._CLIENT_ID} + if secret_token: + query['secret_token'] = secret_token + + if not extract_flat and info.get('downloadable') and info.get('has_downloads_left'): + download_url = update_url_query( + self._API_V2_BASE + 'tracks/' + track_id + '/download', query) + redirect_url = (self._download_json(download_url, track_id, fatal=False) or {}).get('redirectUri') + if redirect_url: + urlh = self._request_webpage( + HEADRequest(redirect_url), track_id, fatal=False) + if urlh: + format_url = urlh.geturl() + format_urls.add(format_url) + formats.append({ + 'format_id': 'download', + 'ext': urlhandle_detect_ext(urlh) or 'mp3', + 'filesize': int_or_none(urlh.headers.get('Content-Length')), + 'url': format_url, + 'quality': 10, + }) + + def invalid_url(url): + return not url or url in format_urls + + def add_format(f, protocol, is_preview=False): + mobj = re.search(r'\.(?P<abr>\d+)\.(?P<ext>[0-9a-z]{3,4})(?=[/?])', stream_url) + if mobj: + for k, v in mobj.groupdict().items(): + if not f.get(k): + f[k] = v + format_id_list = [] + if protocol: + format_id_list.append(protocol) + ext = f.get('ext') + if ext == 'aac': + f['abr'] = '256' + for k in ('ext', 'abr'): + v = f.get(k) + if v: + format_id_list.append(v) + preview = is_preview or re.search(r'/(?:preview|playlist)/0/30/', f['url']) + if preview: + format_id_list.append('preview') + abr = f.get('abr') + if abr: + f['abr'] = int(abr) + if protocol == 'hls': + protocol = 'm3u8' if ext == 'aac' else 'm3u8_native' + else: + protocol = 'http' + f.update({ + 'format_id': '_'.join(format_id_list), + 'protocol': protocol, + 'preference': -10 if preview else None, + }) + formats.append(f) + + # New API + transcodings = try_get( + info, lambda x: x['media']['transcodings'], list) or [] + for t in transcodings: + if not isinstance(t, dict): + continue + format_url = url_or_none(t.get('url')) + if not format_url: + continue + stream = None if extract_flat else self._download_json( + format_url, track_id, query=query, fatal=False, headers=self._HEADERS) + if not isinstance(stream, dict): + continue + stream_url = url_or_none(stream.get('url')) + if invalid_url(stream_url): + continue + format_urls.add(stream_url) + stream_format = t.get('format') or {} + protocol = stream_format.get('protocol') + if protocol != 'hls' and '/hls' in format_url: + protocol = 'hls' + ext = None + preset = str_or_none(t.get('preset')) + if preset: + ext = preset.split('_')[0] + if ext not in KNOWN_EXTENSIONS: + ext = mimetype2ext(stream_format.get('mime_type')) + add_format({ + 'url': stream_url, + 'ext': ext, + }, 'http' if protocol == 'progressive' else protocol, + t.get('snipped') or '/preview/' in format_url) + + for f in formats: + f['vcodec'] = 'none' + + if not formats and info.get('policy') == 'BLOCK': + self.raise_geo_restricted(metadata_available=True) + + user = info.get('user') or {} + + thumbnails = [] + artwork_url = info.get('artwork_url') + thumbnail = artwork_url or user.get('avatar_url') + if isinstance(thumbnail, compat_str): + if re.search(self._IMAGE_REPL_RE, thumbnail): + for image_id, size in self._ARTWORK_MAP.items(): + i = { + 'id': image_id, + 'url': re.sub(self._IMAGE_REPL_RE, '-%s.jpg' % image_id, thumbnail), + } + if image_id == 'tiny' and not artwork_url: + size = 18 + elif image_id == 'original': + i['preference'] = 10 + if size: + i.update({ + 'width': size, + 'height': size, + }) + thumbnails.append(i) + else: + thumbnails = [{'url': thumbnail}] + + def extract_count(key): + return int_or_none(info.get('%s_count' % key)) + + return { + 'id': track_id, + 'uploader': user.get('username'), + 'uploader_id': str_or_none(user.get('id')) or user.get('permalink'), + 'uploader_url': user.get('permalink_url'), + 'timestamp': unified_timestamp(info.get('created_at')), + 'title': title, + 'description': info.get('description'), + 'thumbnails': thumbnails, + 'duration': float_or_none(info.get('duration'), 1000), + 'webpage_url': info.get('permalink_url'), + 'license': info.get('license'), + 'view_count': extract_count('playback'), + 'like_count': extract_count('favoritings') or extract_count('likes'), + 'comment_count': extract_count('comment'), + 'repost_count': extract_count('reposts'), + 'genre': info.get('genre'), + 'formats': formats if not extract_flat else None + } + @classmethod def _resolv_url(cls, url): return cls._API_V2_BASE + 'resolve?url=' + url @@ -387,173 +543,6 @@ class SoundcloudIE(SoundcloudBaseIE): }, ] - _IMAGE_REPL_RE = r'-([0-9a-z]+)\.jpg' - - _ARTWORK_MAP = { - 'mini': 16, - 'tiny': 20, - 'small': 32, - 'badge': 47, - 't67x67': 67, - 'large': 100, - 't300x300': 300, - 'crop': 400, - 't500x500': 500, - 'original': 0, - } - - def _extract_info_dict(self, info, full_title=None, secret_token=None): - track_id = compat_str(info['id']) - title = info['title'] - - format_urls = set() - formats = [] - query = {'client_id': self._CLIENT_ID} - if secret_token: - query['secret_token'] = secret_token - - if info.get('downloadable') and info.get('has_downloads_left'): - download_url = update_url_query( - self._API_V2_BASE + 'tracks/' + track_id + '/download', query) - redirect_url = (self._download_json(download_url, track_id, fatal=False) or {}).get('redirectUri') - if redirect_url: - urlh = self._request_webpage( - HEADRequest(redirect_url), track_id, fatal=False) - if urlh: - format_url = urlh.geturl() - format_urls.add(format_url) - formats.append({ - 'format_id': 'download', - 'ext': urlhandle_detect_ext(urlh) or 'mp3', - 'filesize': int_or_none(urlh.headers.get('Content-Length')), - 'url': format_url, - 'quality': 10, - }) - - def invalid_url(url): - return not url or url in format_urls - - def add_format(f, protocol, is_preview=False): - mobj = re.search(r'\.(?P<abr>\d+)\.(?P<ext>[0-9a-z]{3,4})(?=[/?])', stream_url) - if mobj: - for k, v in mobj.groupdict().items(): - if not f.get(k): - f[k] = v - format_id_list = [] - if protocol: - format_id_list.append(protocol) - ext = f.get('ext') - if ext == 'aac': - f['abr'] = '256' - for k in ('ext', 'abr'): - v = f.get(k) - if v: - format_id_list.append(v) - preview = is_preview or re.search(r'/(?:preview|playlist)/0/30/', f['url']) - if preview: - format_id_list.append('preview') - abr = f.get('abr') - if abr: - f['abr'] = int(abr) - if protocol == 'hls': - protocol = 'm3u8' if ext == 'aac' else 'm3u8_native' - else: - protocol = 'http' - f.update({ - 'format_id': '_'.join(format_id_list), - 'protocol': protocol, - 'preference': -10 if preview else None, - }) - formats.append(f) - - # New API - transcodings = try_get( - info, lambda x: x['media']['transcodings'], list) or [] - for t in transcodings: - if not isinstance(t, dict): - continue - format_url = url_or_none(t.get('url')) - if not format_url: - continue - stream = self._download_json( - format_url, track_id, query=query, fatal=False, headers=self._HEADERS) - if not isinstance(stream, dict): - continue - stream_url = url_or_none(stream.get('url')) - if invalid_url(stream_url): - continue - format_urls.add(stream_url) - stream_format = t.get('format') or {} - protocol = stream_format.get('protocol') - if protocol != 'hls' and '/hls' in format_url: - protocol = 'hls' - ext = None - preset = str_or_none(t.get('preset')) - if preset: - ext = preset.split('_')[0] - if ext not in KNOWN_EXTENSIONS: - ext = mimetype2ext(stream_format.get('mime_type')) - add_format({ - 'url': stream_url, - 'ext': ext, - }, 'http' if protocol == 'progressive' else protocol, - t.get('snipped') or '/preview/' in format_url) - - for f in formats: - f['vcodec'] = 'none' - - if not formats and info.get('policy') == 'BLOCK': - self.raise_geo_restricted(metadata_available=True) - self._sort_formats(formats) - - user = info.get('user') or {} - - thumbnails = [] - artwork_url = info.get('artwork_url') - thumbnail = artwork_url or user.get('avatar_url') - if isinstance(thumbnail, compat_str): - if re.search(self._IMAGE_REPL_RE, thumbnail): - for image_id, size in self._ARTWORK_MAP.items(): - i = { - 'id': image_id, - 'url': re.sub(self._IMAGE_REPL_RE, '-%s.jpg' % image_id, thumbnail), - } - if image_id == 'tiny' and not artwork_url: - size = 18 - elif image_id == 'original': - i['preference'] = 10 - if size: - i.update({ - 'width': size, - 'height': size, - }) - thumbnails.append(i) - else: - thumbnails = [{'url': thumbnail}] - - def extract_count(key): - return int_or_none(info.get('%s_count' % key)) - - return { - 'id': track_id, - 'uploader': user.get('username'), - 'uploader_id': str_or_none(user.get('id')) or user.get('permalink'), - 'uploader_url': user.get('permalink_url'), - 'timestamp': unified_timestamp(info.get('created_at')), - 'title': title, - 'description': info.get('description'), - 'thumbnails': thumbnails, - 'duration': float_or_none(info.get('duration'), 1000), - 'webpage_url': info.get('permalink_url'), - 'license': info.get('license'), - 'view_count': extract_count('playback'), - 'like_count': extract_count('favoritings') or extract_count('likes'), - 'comment_count': extract_count('comment'), - 'repost_count': extract_count('reposts'), - 'genre': info.get('genre'), - 'formats': formats - } - def _real_extract(self, url): mobj = self._match_valid_url(url) @@ -670,25 +659,20 @@ class SoundcloudPagedPlaylistBaseIE(SoundcloudBaseIE): 'offset': 0, } - retries = self.get_param('extractor_retries', 3) - for i in itertools.count(): - attempt, last_error = -1, None - while attempt < retries: - attempt += 1 - if last_error: - self.report_warning('%s. Retrying ...' % remove_end(last_error, '.'), playlist_id) + for retry in self.RetryManager(): try: response = self._download_json( url, playlist_id, query=query, headers=self._HEADERS, - note='Downloading track page %s%s' % (i + 1, f' (retry #{attempt})' if attempt else '')) + note=f'Downloading track page {i + 1}') break except ExtractorError as e: # Downloading page may result in intermittent 502 HTTP error # See https://github.com/hypervideo/hypervideo/issues/872 - if attempt >= retries or not isinstance(e.cause, compat_HTTPError) or e.cause.code != 502: + if not isinstance(e.cause, compat_HTTPError) or e.cause.code != 502: raise - last_error = str(e.cause or e.msg) + retry.error = e + continue def resolve_entry(*candidates): for cand in candidates: @@ -906,6 +890,7 @@ class SoundcloudSearchIE(SoundcloudBaseIE, SearchInfoExtractor): _TESTS = [{ 'url': 'scsearch15:post-avant jazzcore', 'info_dict': { + 'id': 'post-avant jazzcore', 'title': 'post-avant jazzcore', }, 'playlist_count': 15, @@ -932,7 +917,8 @@ class SoundcloudSearchIE(SoundcloudBaseIE, SearchInfoExtractor): for item in response.get('collection') or []: if item: - yield self.url_result(item['uri'], SoundcloudIE.ie_key()) + yield self.url_result( + item['uri'], SoundcloudIE.ie_key(), **self._extract_info_dict(item, extract_flat=True)) next_url = response.get('next_href') if not next_url: |