Merge branch 'master' into HEAD

youtube_dl/extractor/__init__.py

@@ -359,6 +359,7 @@ from .nbc import (
     NBCNewsIE,
     NBCSportsIE,
     NBCSportsVPlayerIE,
+    MSNBCIE,
 )
 from .ndr import (
     NDRIE,
@@ -613,7 +614,10 @@ from .testurl import TestURLIE
 from .testtube import TestTubeIE
 from .tf1 import TF1IE
 from .theonion import TheOnionIE
-from .theplatform import ThePlatformIE
+from .theplatform import (
+    ThePlatformIE,
+    ThePlatformFeedIE,
+)
 from .thesixtyone import TheSixtyOneIE
 from .thisamericanlife import ThisAmericanLifeIE
 from .thisav import ThisAVIE

youtube_dl/extractor/common.py

@@ -1052,7 +1052,7 @@ class InfoExtractor(object):
         return self._search_regex(
             r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
 
-    def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None):
+    def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
         base = smil_url
         for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
             b = meta.get('base') or meta.get('httpBase')
@@ -1091,6 +1091,12 @@ class InfoExtractor(object):
                     'width': width,
                     'height': height,
                 })
+                if transform_rtmp_url:
+                    streamer, src = transform_rtmp_url(streamer, src)
+                    formats[-1].update({
+                        'url': streamer,
+                        'play_path': src,
+                    })
                 continue
 
             src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)

youtube_dl/extractor/nbc.py

@@ -236,3 +236,28 @@ class NBCNewsIE(InfoExtractor):
                 'url': info['videoAssets'][-1]['publicUrl'],
                 'ie_key': 'ThePlatform',
             }
+
+
+class MSNBCIE(InfoExtractor):
+    # https URLs redirect to corresponding http ones
+    _VALID_URL = r'http://www\.msnbc\.com/[^/]+/watch/(?P<id>[^/]+)'
+    _TEST = {
+        'url': 'http://www.msnbc.com/all-in-with-chris-hayes/watch/the-chaotic-gop-immigration-vote-314487875924',
+        'md5': '6d236bf4f3dddc226633ce6e2c3f814d',
+        'info_dict': {
+            'id': 'n_hayes_Aimm_140801_272214',
+            'ext': 'mp4',
+            'title': 'The chaotic GOP immigration vote',
+            'description': 'The Republican House votes on a border bill that has no chance of getting through the Senate or signed by the President and is drawing criticism from all sides.',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'timestamp': 1406937606,
+            'upload_date': '20140802',
+            'categories': ['MSNBC/Topics/Franchise/Best of last night', 'MSNBC/Topics/General/Congress'],
+        },
+    }
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        embed_url = self._html_search_meta('embedURL', webpage)
+        return self.url_result(embed_url)

youtube_dl/extractor/theplatform.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 from __future__ import unicode_literals
 
 import re
@@ -10,7 +11,8 @@ import hashlib
 
 from .common import InfoExtractor
 from ..compat import (
-    compat_str,
+    compat_parse_qs,
+    compat_urllib_parse_urlparse,
 )
 from ..utils import (
     determine_ext,
@@ -18,12 +20,68 @@ 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_formats(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)
+
+        return formats
+
+    def get_metadata(self, path, video_id):
+        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)
+
+        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<provider_id>[^/]+)/
            (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
@@ -67,6 +125,20 @@ class ThePlatformIE(InfoExtractor):
     }, {
         '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"],
+        },
     }]
 
     @staticmethod
@@ -101,6 +173,24 @@ class ThePlatformIE(InfoExtractor):
             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'<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(script, 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
         elif mobj.group('config'):
@@ -120,95 +210,78 @@ class ThePlatformIE(InfoExtractor):
         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 = self._extract_theplatform_smil_formats(smil_url, video_id)
 
-        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)
-
-        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)
-
-        return {
+        ret = self.get_metadata(path, video_id)
+        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),
-        }
+        })
+
+        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<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[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$',
+            'upload_date': '20140208',
+            'timestamp': 1391824260,
+            'duration': 467.0,
+            'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
+        },
+    }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+
+        video_id = mobj.group('id')
+        provider_id = mobj.group('provider_id')
+        feed_id = mobj.group('feed_id')
+
+        real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
+        feed = self._download_json(real_url, video_id)
+        entry = feed['entries'][0]
+
+        formats = []
+        first_video_id = None
+        duration = None
+        for item in entry['media$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'))
+            formats.extend(self._extract_theplatform_smil_formats(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id))
+
+        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.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
+        ret.update({
+            'id': video_id,
+            'formats': formats,
+            'thumbnails': thumbnails,
+            'duration': duration,
+            'timestamp': timestamp,
+            'categories': categories,
+        })
+
+        return ret