youtube-dl/youtube_dl/extractor/theplatform.py

376 lines
15 KiB
Python
Raw Normal View History

2016-10-02 05:39:18 -06:00
# coding: utf-8
from __future__ import unicode_literals
2013-12-04 15:28:40 -07:00
import re
2015-02-14 15:56:12 -07:00
import time
import hmac
import binascii
import hashlib
2013-12-04 15:28:40 -07:00
from .once import OnceIE
from .adobepass import AdobePassIE
from ..compat import (
compat_parse_qs,
compat_urllib_parse_urlparse,
)
from ..utils import (
determine_ext,
ExtractorError,
float_or_none,
2015-04-08 10:29:10 -06:00
int_or_none,
sanitized_Request,
unsmuggle_url,
update_url_query,
xpath_with_ns,
mimetype2ext,
find_xpath_attr,
2013-12-04 15:28:40 -07:00
)
default_ns = 'http://www.w3.org/2005/SMIL21/Language'
_x = lambda p: xpath_with_ns(p, {'smil': default_ns})
2013-12-04 15:28:40 -07:00
class ThePlatformBaseIE(OnceIE):
def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
meta = self._download_xml(
smil_url, video_id, note=note, query={'format': 'SMIL'},
headers=self.geo_verification_headers())
error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
if error_element is not None and error_element.attrib['src'].startswith(
'http://link.theplatform.com/s/errorFiles/Unavailable.'):
raise ExtractorError(error_element.attrib['abstract'], expected=True)
2015-08-19 11:24:32 -06:00
smil_formats = self._parse_smil_formats(
2015-08-19 11:24:32 -06:00
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))
formats = []
for _format in smil_formats:
if OnceIE.suitable(_format['url']):
formats.extend(self._extract_once_formats(_format['url']))
else:
media_url = _format['url']
if determine_ext(media_url) == 'm3u8':
hdnea2 = self._get_cookies(media_url).get('hdnea2')
if hdnea2:
_format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
formats.append(_format)
2015-08-19 11:24:32 -06:00
subtitles = self._parse_smil_subtitles(meta, default_ns)
return formats, subtitles
2015-08-19 11:24:32 -06:00
def _download_theplatform_metadata(self, path, video_id):
2015-08-19 11:24:32 -06:00
info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
return self._download_json(info_url, video_id)
2015-08-19 11:24:32 -06:00
def _parse_theplatform_metadata(self, info):
2015-08-19 11:24:32 -06:00
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.setdefault(lang, []).append({
'ext': mimetype2ext(mime),
2015-08-19 11:24:32 -06:00
'url': src,
})
2015-08-19 11:24:32 -06:00
return {
'title': info['title'],
'subtitles': subtitles,
'description': info['description'],
'thumbnail': info['defaultThumbnailUrl'],
'duration': int_or_none(info.get('duration'), 1000),
'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
'uploader': info.get('billingCode'),
2015-08-19 11:24:32 -06:00
}
def _extract_theplatform_metadata(self, path, video_id):
info = self._download_theplatform_metadata(path, video_id)
return self._parse_theplatform_metadata(info)
2015-08-19 11:24:32 -06:00
class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
_VALID_URL = r'''(?x)
2015-02-14 15:56:12 -07:00
(?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
(?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
|theplatform:)(?P<id>[^/\?&]+)'''
2013-12-04 15:28:40 -07:00
_TESTS = [{
2013-12-04 15:28:40 -07:00
# from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
'info_dict': {
'id': 'e9I_cZgTgIPd',
'ext': 'flv',
'title': 'Blackberry\'s big, bold Z30',
'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
'duration': 247,
'timestamp': 1383239700,
'upload_date': '20131031',
'uploader': 'CBSI-NEW',
2013-12-04 15:28:40 -07:00
},
'params': {
2013-12-04 15:28:40 -07:00
# rtmp download
'skip_download': True,
2013-12-04 15:28:40 -07:00
},
'skip': '404 Not Found',
}, {
# from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
'info_dict': {
'id': '22d_qsQ6MIRT',
'ext': 'flv',
'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
'title': 'Tesla Model S: A second step towards a cleaner motoring future',
'timestamp': 1426176191,
'upload_date': '20150312',
'uploader': 'CBSI-NEW',
},
'params': {
# 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',
'uploader': 'EGSM',
}
}, {
'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': 'fb96bb3d85118930a5b055783a3bd992',
'info_dict': {
'id': 'tdy_or_siri_150701',
'ext': 'mp4',
'title': 'iPhone Siris sassy response to a math question has people talking',
'description': 'md5:a565d1deadd5086f3331d57298ec6333',
'duration': 83.0,
'thumbnail': 're:^https?://.*\.jpg$',
'timestamp': 1435752600,
'upload_date': '20150701',
'uploader': 'NBCU-NEWS',
},
2015-11-15 11:28:04 -07:00
}, {
# 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,
}]
2014-11-23 12:41:03 -07:00
2016-05-21 18:47:22 -06:00
@classmethod
def _extract_urls(cls, webpage):
m = re.search(
r'''(?x)
<meta\s+
property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
''', webpage)
if m:
return [m.group('url')]
matches = re.findall(
r'<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
if matches:
return list(zip(*matches))[1]
2015-02-14 15:56:12 -07:00
@staticmethod
def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
flags = '10' if include_qs else '00'
expiration_date = '%x' % (int(time.time()) + life)
def str_to_hex(str):
return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
def hex_to_bytes(hex):
return binascii.a2b_hex(hex.encode('ascii'))
2015-02-14 15:56:12 -07:00
2016-04-01 02:55:46 -06:00
relative_path = re.match(r'https?://link.theplatform.com/s/([^?]+)', url).group(1)
clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
2015-02-14 15:56:12 -07:00
checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
return '%s&sig=%s' % (url, sig)
2014-10-26 18:14:17 -06:00
def _real_extract(self, url):
2015-02-14 15:56:12 -07:00
url, smuggled_data = unsmuggle_url(url, {})
2014-10-26 18:14:17 -06:00
mobj = re.match(self._VALID_URL, url)
2015-02-14 15:56:12 -07:00
provider_id = mobj.group('provider_id')
2014-10-26 18:14:17 -06:00
video_id = mobj.group('id')
2015-02-14 15:56:12 -07:00
if not provider_id:
provider_id = 'dJ5BDC'
2016-03-30 16:18:31 -06:00
path = provider_id + '/'
if mobj.group('media'):
2016-03-30 16:18:31 -06:00
path += mobj.group('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'<script[^>]+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(
2015-11-19 09:58:43 -07:00
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:
headers = {}
source_url = smuggled_data.get('source_url')
if source_url:
headers['Referer'] = source_url
request = sanitized_Request(url, headers=headers)
webpage = self._download_webpage(request, video_id)
smil_url = self._search_regex(
r'<link[^>]+href=(["\'])(?P<url>.+?)\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'
elif mobj.group('config'):
2014-11-23 12:41:03 -07:00
config_url = url + '&form=json'
2014-10-26 18:14:17 -06:00
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')
if 'releaseUrl' in config:
release_url = config['releaseUrl']
else:
release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
smil_url = release_url + '&formats=MPEG4&manifest=f4m'
2014-10-26 18:14:17 -06:00
else:
smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
2015-02-14 15:56:12 -07:00
sig = smuggled_data.get('sig')
if sig:
smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
2013-12-04 15:28:40 -07:00
formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
self._sort_formats(formats)
ret = self._extract_theplatform_metadata(path, video_id)
combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
2015-08-19 11:24:32 -06:00
ret.update({
'id': video_id,
'formats': formats,
'subtitles': combined_subtitles,
2015-08-19 11:24:32 -06:00
})
2013-12-04 15:28:40 -07:00
2015-08-19 11:24:32 -06:00
return ret
2015-08-19 11:24:32 -06:00
class ThePlatformFeedIE(ThePlatformBaseIE):
_URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
_VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[\w-]+))'
_TESTS = [{
2015-08-19 11:24:32 -06:00
# 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': '6e32495b5073ab414471b615c5ded394',
2015-08-19 11:24:32 -06:00
'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'],
'uploader': 'NBCU-NEWS',
2015-08-19 11:24:32 -06:00
},
}]
2015-08-19 11:24:32 -06:00
def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}):
real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
entry = self._download_json(real_url, video_id)['entries'][0]
2015-08-19 11:24:32 -06:00
formats = []
subtitles = {}
2015-08-19 11:24:32 -06:00
first_video_id = None
duration = None
asset_types = []
2015-08-19 11:24:32 -06:00
for item in entry['media$content']:
smil_url = item['plfile$url']
cur_video_id = ThePlatformIE._match_id(smil_url)
2015-08-19 11:24:32 -06:00
if first_video_id is None:
first_video_id = cur_video_id
duration = float_or_none(item.get('plfile$duration'))
for asset_type in item['plfile$assetTypes']:
if asset_type in asset_types:
continue
asset_types.append(asset_type)
query = {
'mbr': 'true',
'formats': item['plfile$format'],
'assetTypes': asset_type,
}
if asset_type in asset_types_query:
query.update(asset_types_query[asset_type])
cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
formats.extend(cur_formats)
subtitles = self._merge_subtitles(subtitles, cur_subtitles)
self._sort_formats(formats)
2013-12-04 15:28:40 -07:00
2015-08-19 11:24:32 -06:00
thumbnails = [{
'url': thumbnail['plfile$url'],
'width': int_or_none(thumbnail.get('plfile$width')),
'height': int_or_none(thumbnail.get('plfile$height')),
} for thumbnail in entry.get('media$thumbnails', [])]
timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
categories = [item['media$name'] for item in entry.get('media$categories', [])]
ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
2015-08-19 11:24:32 -06:00
ret.update({
2013-12-04 15:28:40 -07:00
'id': video_id,
'formats': formats,
'subtitles': subtitles,
2015-08-19 11:24:32 -06:00
'thumbnails': thumbnails,
'duration': duration,
'timestamp': timestamp,
'categories': categories,
})
if custom_fields:
ret.update(custom_fields(entry))
2015-08-19 11:24:32 -06:00
return ret
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')
filter_query = mobj.group('filter')
return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)