Prepare for upload.

[youtubedl] / youtube_dl / extractor / theplatform.py

diff --git a/youtube_dl/extractor/theplatform.py b/youtube_dl/extractor/theplatform.py
index 83d833e30dbeb60caa43aa272bfd4d35f4507a53..9a424b1c6aeb089af8050d7eee6b29591968c3aa 100644 (file)
--- a/youtube_dl/extractor/theplatform.py
+++ b/youtube_dl/extractor/theplatform.py
@@ -1,32 +1,104 @@
+# coding: utf-8

 from __future__ import unicode_literals
 
 import re
 from __future__ import unicode_literals
 
 import re

-import json

 import time
 import hmac
 import binascii
 import hashlib
 
 
 import time
 import hmac
 import binascii
 import hashlib
 
 

-from .common import InfoExtractor
+from .once import OnceIE
+from .adobepass import AdobePassIE

 from ..compat import (
 from ..compat import (

-    compat_str,
+    compat_parse_qs,
+    compat_urllib_parse_urlparse,

 )
 from ..utils import (
     determine_ext,
     ExtractorError,
 )
 from ..utils import (
     determine_ext,
     ExtractorError,

-    xpath_with_ns,
-    unsmuggle_url,
+    float_or_none,

     int_or_none,
     int_or_none,

+    sanitized_Request,
+    unsmuggle_url,
+    update_url_query,
+    xpath_with_ns,
+    mimetype2ext,
+    find_xpath_attr,

 )
 
 )
 

-_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 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)
+
+        smil_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))
+
+        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)
+
+        subtitles = self._parse_smil_subtitles(meta, default_ns)
+
+        return formats, subtitles
+
+    def _download_theplatform_metadata(self, path, video_id):
+        info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
+        return self._download_json(info_url, video_id)
+
+    def _parse_theplatform_metadata(self, info):
+        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),
+                    'url': src,
+                })
+
+        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'),
+        }
+
+    def _extract_theplatform_metadata(self, path, video_id):
+        info = self._download_theplatform_metadata(path, video_id)
+        return self._parse_theplatform_metadata(info)

 
 
 
 

-class ThePlatformIE(InfoExtractor):
+class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):

     _VALID_URL = r'''(?x)
         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
     _VALID_URL = r'''(?x)
         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/

-           (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
+           (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?

          |theplatform:)(?P<id>[^/\?&]+)'''
 
     _TESTS = [{
          |theplatform:)(?P<id>[^/\?&]+)'''
 
     _TESTS = [{


@@ -38,11 +110,15 @@ class ThePlatformIE(InfoExtractor):
             'title': 'Blackberry\'s big, bold Z30',
             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
             'duration': 247,
             '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',

         },
         'params': {
             # rtmp download
             'skip_download': True,
         },
         },
         'params': {
             # rtmp download
             'skip_download': True,
         },

+        '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',
     }, {
         # 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',


@@ -51,6 +127,9 @@ class ThePlatformIE(InfoExtractor):
             'ext': 'flv',
             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
             '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
         },
         'params': {
             # rtmp download


@@ -63,12 +142,50 @@ class ThePlatformIE(InfoExtractor):
             'ext': 'mp4',
             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
             '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/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 Siri’s sassy response to a math question has people talking',
+            'description': 'md5:a565d1deadd5086f3331d57298ec6333',
+            'duration': 83.0,
+            'thumbnail': r're:^https?://.*\.jpg$',
+            'timestamp': 1435752600,
+            'upload_date': '20150701',
+            'uploader': 'NBCU-NEWS',
+        },
+    }, {
+        # 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,

     }]
 
     }]
 

+    @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')]
+
+        # Are whitesapces ignored in URLs?
+        # https://github.com/rg3/youtube-dl/issues/12044
+        matches = re.findall(
+            r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
+        if matches:
+            return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
+

     @staticmethod
     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
         flags = '10' if include_qs else '00'
     @staticmethod
     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
         flags = '10' if include_qs else '00'


@@ -77,11 +194,11 @@ class ThePlatformIE(InfoExtractor):
         def str_to_hex(str):
             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 
         def str_to_hex(str):
             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 

-        def hex_to_str(hex):
-            return binascii.a2b_hex(hex)
+        def hex_to_bytes(hex):
+            return binascii.a2b_hex(hex.encode('ascii'))

 
 

-        relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
-        clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
+        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))

         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)
         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)


@@ -96,115 +213,166 @@ class ThePlatformIE(InfoExtractor):
         if not provider_id:
             provider_id = 'dJ5BDC'
 
         if not provider_id:
             provider_id = 'dJ5BDC'
 

-        path = provider_id
+        path = provider_id + '/'

         if mobj.group('media'):
         if mobj.group('media'):

-            path += '/media'
-        path += '/' + video_id
+            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(
+                    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
 
         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'):
             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')
         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 + '&formats=MPEG4&manifest=f4m'

         else:
         else:

-            smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
+            smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path

 
         sig = smuggled_data.get('sig')
         if sig:
             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 
 
         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)
+        self._sort_formats(formats)
+
+        ret = self._extract_theplatform_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

 
 

-        info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
-        info_json = self._download_webpage(info_url, video_id)
-        info = json.loads(info_json)

 
 

+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 = [{
+        # 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',
+        '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': r're:^https?://.*\.jpg$',
+            'upload_date': '20140208',
+            'timestamp': 1391824260,
+            'duration': 467.0,
+            'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
+            'uploader': 'NBCU-NEWS',
+        },
+    }]
+
+    def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
+        real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
+        entry = self._download_json(real_url, video_id)['entries'][0]
+        main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else None
+
+        formats = []

         subtitles = {}
         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,
-                }]
-
-        head = meta.find(_x('smil:head'))
-        body = meta.find(_x('smil:body'))
-
-        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)
+        first_video_id = None
+        duration = None
+        asset_types = []
+        for item in entry['media$content']:
+            smil_url = item['plfile$url']
+            cur_video_id = ThePlatformIE._match_id(smil_url)
+            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(
+                    main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
+                formats.extend(cur_formats)
+                subtitles = self._merge_subtitles(subtitles, cur_subtitles)

 
 

-        return {
+        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', [])]
+
+        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'])
+        ret.update({

             'id': video_id,
             'id': video_id,

-            'title': info['title'],
-            'subtitles': subtitles,

             'formats': formats,
             '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,
+        })
+        if custom_fields:
+            ret.update(custom_fields(entry))
+
+        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)