X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/540fd68c40df72763aee5d75598675c45cfa9aba..d018d3313032e12968a6add6800e51d412e2f602:/youtube_dl/extractor/theplatform.py?ds=inline diff --git a/youtube_dl/extractor/theplatform.py b/youtube_dl/extractor/theplatform.py index 92731ad..1555aa7 100644 --- a/youtube_dl/extractor/theplatform.py +++ b/youtube_dl/extractor/theplatform.py @@ -1,7 +1,7 @@ +# -*- coding: utf-8 -*- from __future__ import unicode_literals import re -import json import time import hmac import binascii @@ -10,7 +10,8 @@ import hashlib from .common import InfoExtractor from ..compat import ( - compat_str, + compat_parse_qs, + compat_urllib_parse_urlparse, ) from ..utils import ( determine_ext, @@ -18,15 +19,72 @@ from ..utils import ( xpath_with_ns, unsmuggle_url, int_or_none, + url_basename, + float_or_none, ) -_x = lambda p: xpath_with_ns(p, {'smil': 'http://www.w3.org/2005/SMIL21/Language'}) +default_ns = 'http://www.w3.org/2005/SMIL21/Language' +_x = lambda p: xpath_with_ns(p, {'smil': default_ns}) -class ThePlatformIE(InfoExtractor): +class ThePlatformBaseIE(InfoExtractor): + def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'): + meta = self._download_xml(smil_url, video_id, note=note) + try: + error_msg = next( + n.attrib['abstract'] + for n in meta.findall(_x('.//smil:ref')) + if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired') + except StopIteration: + pass + else: + raise ExtractorError(error_msg, expected=True) + + formats = self._parse_smil_formats( + meta, smil_url, video_id, namespace=default_ns, + # the parameters are from syfy.com, other sites may use others, + # they also work for nbc.com + f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'}, + transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src)) + + for _format in formats: + ext = determine_ext(_format['url']) + if ext == 'once': + _format['ext'] = 'mp4' + + self._sort_formats(formats) + + subtitles = self._parse_smil_subtitles(meta, default_ns) + + return formats, subtitles + + def get_metadata(self, path, video_id): + info_url = 'http://link.theplatform.com/s/%s?format=preview' % path + info = self._download_json(info_url, video_id) + + subtitles = {} + captions = info.get('captions') + if isinstance(captions, list): + for caption in captions: + lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type') + subtitles[lang] = [{ + 'ext': 'srt' if mime == 'text/srt' else 'ttml', + 'url': src, + }] + + return { + 'title': info['title'], + 'subtitles': subtitles, + 'description': info['description'], + 'thumbnail': info['defaultThumbnailUrl'], + 'duration': int_or_none(info.get('duration'), 1000), + } + + +class ThePlatformIE(ThePlatformBaseIE): _VALID_URL = r'''(?x) (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P[^/]+)/ - (?P(?:[^/\?]+/(?:swf|config)|onsite)/select/)? + (?:(?P(?:[^/]+/)+select/media/)|(?P(?:[^/\?]+/(?:swf|config)|onsite)/select/))? |theplatform:)(?P[^/\?&]+)''' _TESTS = [{ @@ -56,6 +114,36 @@ class ThePlatformIE(InfoExtractor): # rtmp download 'skip_download': True, } + }, { + 'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD', + 'info_dict': { + 'id': 'yMBg9E8KFxZD', + 'ext': 'mp4', + 'description': 'md5:644ad9188d655b742f942bf2e06b002d', + 'title': 'HIGHLIGHTS: USA bag first ever series Cup win', + } + }, { + 'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7', + 'only_matching': True, + }, { + 'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701', + 'md5': '734f3790fb5fc4903da391beeebc4836', + 'info_dict': { + 'id': 'tdy_or_siri_150701', + 'ext': 'mp4', + 'title': 'iPhone Siri’s sassy response to a math question has people talking', + 'description': 'md5:a565d1deadd5086f3331d57298ec6333', + 'duration': 83.0, + 'thumbnail': 're:^https?://.*\.jpg$', + 'timestamp': 1435752600, + 'upload_date': '20150701', + 'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"], + }, + }, { + # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1 + # geo-restricted (US), HLS encrypted with AES-128 + 'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781', + 'only_matching': True, }] @staticmethod @@ -85,111 +173,140 @@ class ThePlatformIE(InfoExtractor): if not provider_id: provider_id = 'dJ5BDC' + path = provider_id + if mobj.group('media'): + path += '/media' + path += '/' + video_id + + qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query) + if 'guid' in qs_dict: + webpage = self._download_webpage(url, video_id) + scripts = re.findall(r']+src="([^"]+)"', webpage) + feed_id = None + # feed id usually locates in the last script. + # Seems there's no pattern for the interested script filename, so + # I try one by one + for script in reversed(scripts): + feed_script = self._download_webpage( + self._proto_relative_url(script, 'http:'), + video_id, 'Downloading feed script') + feed_id = self._search_regex( + r'defaultFeedId\s*:\s*"([^"]+)"', feed_script, + 'default feed id', default=None) + if feed_id is not None: + break + if feed_id is None: + raise ExtractorError('Unable to find feed id') + return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % ( + provider_id, feed_id, qs_dict['guid'][0])) + if smuggled_data.get('force_smil_url', False): smil_url = url + # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385) + elif '/guid/' in url: + webpage = self._download_webpage(url, video_id) + smil_url = self._search_regex( + r']+href=(["\'])(?P.+?)\1[^>]+type=["\']application/smil\+xml', + webpage, 'smil url', group='url') + path = self._search_regex( + r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path') + smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4&format=SMIL' elif mobj.group('config'): config_url = url + '&form=json' config_url = config_url.replace('swf/', 'config/') config_url = config_url.replace('onsite/', 'onsite/config/') config = self._download_json(config_url, video_id, 'Downloading config') - smil_url = config['releaseUrl'] + '&format=SMIL&formats=MPEG4&manifest=f4m' + if 'releaseUrl' in config: + release_url = config['releaseUrl'] + else: + release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path + smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m' else: - smil_url = ('http://link.theplatform.com/s/{0}/{1}/meta.smil?' - 'format=smil&mbr=true'.format(provider_id, video_id)) + smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path sig = smuggled_data.get('sig') if sig: smil_url = self._sign_url(smil_url, sig['key'], sig['secret']) - meta = self._download_xml(smil_url, video_id) - try: - error_msg = next( - n.attrib['abstract'] - for n in meta.findall(_x('.//smil:ref')) - if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired') - except StopIteration: - pass - else: - raise ExtractorError(error_msg, expected=True) + formats, subtitles = self._extract_theplatform_smil(smil_url, video_id) + + ret = self.get_metadata(path, video_id) + combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles) + ret.update({ + 'id': video_id, + 'formats': formats, + 'subtitles': combined_subtitles, + }) + + return ret + + +class ThePlatformFeedIE(ThePlatformBaseIE): + _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s' + _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P[^/]+)/(?P[^?/]+)\?(?:[^&]+&)*byGuid=(?P[a-zA-Z0-9_]+)' + _TEST = { + # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207 + 'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207', + 'md5': '22d2b84f058d3586efcd99e57d59d314', + 'info_dict': { + 'id': 'n_hardball_5biden_140207', + 'ext': 'mp4', + 'title': 'The Biden factor: will Joe run in 2016?', + 'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.', + 'thumbnail': 're:^https?://.*\.jpg$', + 'upload_date': '20140208', + 'timestamp': 1391824260, + 'duration': 467.0, + 'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'], + }, + } + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + + video_id = mobj.group('id') + provider_id = mobj.group('provider_id') + feed_id = mobj.group('feed_id') - info_url = 'http://link.theplatform.com/s/{0}/{1}?format=preview'.format(provider_id, video_id) - info_json = self._download_webpage(info_url, video_id) - info = json.loads(info_json) + real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id) + feed = self._download_json(real_url, video_id) + entry = feed['entries'][0] + formats = [] subtitles = {} - captions = info.get('captions') - if isinstance(captions, list): - for caption in captions: - lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type') - subtitles[lang] = [{ - 'ext': 'srt' if mime == 'text/srt' else 'ttml', - 'url': src, - }] + first_video_id = None + duration = None + for item in entry['media$content']: + smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M' + cur_video_id = url_basename(smil_url) + if first_video_id is None: + first_video_id = cur_video_id + duration = float_or_none(item.get('plfile$duration')) + cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id) + formats.extend(cur_formats) + subtitles = self._merge_subtitles(subtitles, cur_subtitles) - head = meta.find(_x('smil:head')) - body = meta.find(_x('smil:body')) + self._sort_formats(formats) - f4m_node = body.find(_x('smil:seq//smil:video')) - if f4m_node is None: - f4m_node = body.find(_x('smil:seq/smil:video')) - if f4m_node is not None and '.f4m' in f4m_node.attrib['src']: - f4m_url = f4m_node.attrib['src'] - if 'manifest.f4m?' not in f4m_url: - f4m_url += '?' - # the parameters are from syfy.com, other sites may use others, - # they also work for nbc.com - f4m_url += '&g=UXWGVKRWHFSP&hdcore=3.0.3' - formats = self._extract_f4m_formats(f4m_url, video_id) - else: - formats = [] - switch = body.find(_x('smil:switch')) - if switch is None: - switch = body.find(_x('smil:par//smil:switch')) - if switch is None: - switch = body.find(_x('smil:par/smil:switch')) - if switch is None: - switch = body.find(_x('smil:par')) - if switch is not None: - base_url = head.find(_x('smil:meta')).attrib['base'] - for f in switch.findall(_x('smil:video')): - attr = f.attrib - width = int_or_none(attr.get('width')) - height = int_or_none(attr.get('height')) - vbr = int_or_none(attr.get('system-bitrate'), 1000) - format_id = '%dx%d_%dk' % (width, height, vbr) - formats.append({ - 'format_id': format_id, - 'url': base_url, - 'play_path': 'mp4:' + attr['src'], - 'ext': 'flv', - 'width': width, - 'height': height, - 'vbr': vbr, - }) - else: - switch = body.find(_x('smil:seq//smil:switch')) - if switch is None: - switch = body.find(_x('smil:seq/smil:switch')) - for f in switch.findall(_x('smil:video')): - attr = f.attrib - vbr = int_or_none(attr.get('system-bitrate'), 1000) - ext = determine_ext(attr['src']) - if ext == 'once': - ext = 'mp4' - formats.append({ - 'format_id': compat_str(vbr), - 'url': attr['src'], - 'vbr': vbr, - 'ext': ext, - }) - self._sort_formats(formats) + thumbnails = [{ + 'url': thumbnail['plfile$url'], + 'width': int_or_none(thumbnail.get('plfile$width')), + 'height': int_or_none(thumbnail.get('plfile$height')), + } for thumbnail in entry.get('media$thumbnails', [])] - return { + timestamp = int_or_none(entry.get('media$availableDate'), scale=1000) + categories = [item['media$name'] for item in entry.get('media$categories', [])] + + ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id) + subtitles = self._merge_subtitles(subtitles, ret['subtitles']) + ret.update({ 'id': video_id, - 'title': info['title'], - 'subtitles': subtitles, 'formats': formats, - 'description': info['description'], - 'thumbnail': info['defaultThumbnailUrl'], - 'duration': int_or_none(info.get('duration'), 1000), - } + 'subtitles': subtitles, + 'thumbnails': thumbnails, + 'duration': duration, + 'timestamp': timestamp, + 'categories': categories, + }) + + return ret