]> gitweb @ CieloNegro.org - youtube-dl.git/commitdiff
[nbc:nbcnews] improve extraction(fixes #6922)
authorremitamine <remitamine@gmail.com>
Thu, 17 Mar 2016 15:07:36 +0000 (16:07 +0100)
committerremitamine <remitamine@gmail.com>
Thu, 17 Mar 2016 15:11:29 +0000 (16:11 +0100)
- extract more metadata and formats
- relax regex

youtube_dl/extractor/nbc.py

index 2202cfa334765653b44e180306d8e40e0be2049d..bb0817e34ea98edb641580d356b901fee2200c7b 100644 (file)
@@ -3,13 +3,16 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
-from ..compat import compat_HTTPError
+from .theplatform import ThePlatformIE
 from ..utils import (
-    ExtractorError,
     find_xpath_attr,
     lowercase_escape,
     smuggle_url,
     unescapeHTML,
+    update_url_query,
+    int_or_none,
+    HEADRequest,
+    parse_iso8601,
 )
 
 
@@ -131,10 +134,10 @@ class NBCSportsIE(InfoExtractor):
             NBCSportsVPlayerIE._extract_url(webpage), 'NBCSportsVPlayer')
 
 
-class NBCNewsIE(InfoExtractor):
+class NBCNewsIE(ThePlatformIE):
     _VALID_URL = r'''(?x)https?://(?:www\.)?nbcnews\.com/
         (?:video/.+?/(?P<id>\d+)|
-        (?:watch|feature|nightly-news)/[^/]+/(?P<title>.+))
+        ([^/]+/)*(?P<display_id>[^/?]+))
         '''
 
     _TESTS = [
@@ -149,15 +152,14 @@ class NBCNewsIE(InfoExtractor):
             },
         },
         {
-            'url': 'http://www.nbcnews.com/feature/edward-snowden-interview/how-twitter-reacted-snowden-interview-n117236',
-            'md5': 'b2421750c9f260783721d898f4c42063',
+            'url': 'http://www.nbcnews.com/watch/nbcnews-com/how-twitter-reacted-to-the-snowden-interview-269389891880',
+            'md5': 'af1adfa51312291a017720403826bb64',
             'info_dict': {
-                'id': 'I1wpAI_zmhsQ',
+                'id': '269389891880',
                 'ext': 'mp4',
                 'title': 'How Twitter Reacted To The Snowden Interview',
                 'description': 'md5:65a0bd5d76fe114f3c2727aa3a81fe64',
             },
-            'add_ie': ['ThePlatform'],
         },
         {
             'url': 'http://www.nbcnews.com/feature/dateline-full-episodes/full-episode-family-business-n285156',
@@ -168,17 +170,29 @@ class NBCNewsIE(InfoExtractor):
                 'title': 'FULL EPISODE: Family Business',
                 'description': 'md5:757988edbaae9d7be1d585eb5d55cc04',
             },
+            'skip': 'This page is unavailable.',
         },
         {
             'url': 'http://www.nbcnews.com/nightly-news/video/nightly-news-with-brian-williams-full-broadcast-february-4-394064451844',
-            'md5': 'b5dda8cddd8650baa0dcb616dd2cf60d',
+            'md5': '73135a2e0ef819107bbb55a5a9b2a802',
             'info_dict': {
-                'id': 'sekXqyTVnmN3',
+                'id': '394064451844',
                 'ext': 'mp4',
                 'title': 'Nightly News with Brian Williams Full Broadcast (February 4)',
                 'description': 'md5:1c10c1eccbe84a26e5debb4381e2d3c5',
             },
         },
+        {
+            'url': 'http://www.nbcnews.com/business/autos/volkswagen-11-million-vehicles-could-have-suspect-software-emissions-scandal-n431456',
+            'md5': 'a49e173825e5fcd15c13fc297fced39d',
+            'info_dict': {
+                'id': '529953347624',
+                'ext': 'mp4',
+                'title': 'Volkswagen U.S. Chief: We \'Totally Screwed Up\'',
+                'description': 'md5:d22d1281a24f22ea0880741bb4dd6301',
+            },
+            'expected_warnings': ['http-6000 is not available']
+        },
         {
             'url': 'http://www.nbcnews.com/watch/dateline/full-episode--deadly-betrayal-386250819952',
             'only_matching': True,
@@ -202,49 +216,80 @@ class NBCNewsIE(InfoExtractor):
             }
         else:
             # "feature" and "nightly-news" pages use theplatform.com
-            title = mobj.group('title')
-            webpage = self._download_webpage(url, title)
+            display_id = mobj.group('display_id')
+            webpage = self._download_webpage(url, display_id)
+            info = None
             bootstrap_json = self._search_regex(
-                r'var\s+(?:bootstrapJson|playlistData)\s*=\s*({.+});?\s*$',
-                webpage, 'bootstrap json', flags=re.MULTILINE)
-            bootstrap = self._parse_json(bootstrap_json, video_id)
-            info = bootstrap['results'][0]['video']
-            mpxid = info['mpxId']
+                r'(?m)var\s+(?:bootstrapJson|playlistData)\s*=\s*({.+});?\s*$',
+                webpage, 'bootstrap json', default=None)
+            if bootstrap_json:
+                bootstrap = self._parse_json(bootstrap_json, display_id)
+                info = bootstrap['results'][0]['video']
+            else:
+                player_instance_json = self._search_regex(
+                    r'videoObj\s*:\s*({.+})', webpage, 'player instance')
+                info = self._parse_json(player_instance_json, display_id)
+            video_id = info['mpxId']
+            title = info['title']
 
-            base_urls = [
-                info['fallbackPlaylistUrl'],
-                info['associatedPlaylistUrl'],
-            ]
+            subtitles = {}
+            caption_links = info.get('captionLinks')
+            if caption_links:
+                for (sub_key, sub_ext) in (('smpte-tt', 'ttml'), ('web-vtt', 'vtt'), ('srt', 'srt')):
+                    sub_url = caption_links.get(sub_key)
+                    if sub_url:
+                        subtitles.setdefault('en', []).append({
+                            'url': sub_url,
+                            'ext': sub_ext,
+                        })
 
-            for base_url in base_urls:
-                if not base_url:
+            formats = []
+            for video_asset in info['videoAssets']:
+                video_url = video_asset.get('publicUrl')
+                if not video_url:
                     continue
-                playlist_url = base_url + '?form=MPXNBCNewsAPI'
-
-                try:
-                    all_videos = self._download_json(playlist_url, title)
-                except ExtractorError as ee:
-                    if isinstance(ee.cause, compat_HTTPError):
-                        continue
-                    raise
-
-                if not all_videos or 'videos' not in all_videos:
-                    continue
-
-                try:
-                    info = next(v for v in all_videos['videos'] if v['mpxId'] == mpxid)
-                    break
-                except StopIteration:
+                container = video_asset.get('format')
+                asset_type = video_asset.get('assetType') or ''
+                if container == 'ISM' or asset_type == 'FireTV-Once':
                     continue
-
-            if info is None:
-                raise ExtractorError('Could not find video in playlists')
+                elif asset_type == 'OnceURL':
+                    tp_formats, tp_subtitles = self._extract_theplatform_smil(
+                        video_url, video_id)
+                    formats.extend(tp_formats)
+                    subtitles = self._merge_subtitles(subtitles, tp_subtitles)
+                else:
+                    tbr = int_or_none(video_asset.get('bitRate'), 1000)
+                    format_id = 'http%s' % ('-%d' % tbr if tbr else '')
+                    video_url = update_url_query(
+                        video_url, {'format': 'redirect'})
+                    # resolve the url so that we can check availability and detect the correct extension
+                    head = self._request_webpage(
+                        HEADRequest(video_url), video_id,
+                        'Checking %s url' % format_id,
+                        '%s is not available' % format_id,
+                        fatal=False)
+                    if head:
+                        video_url = head.geturl()
+                        formats.append({
+                            'format_id': format_id,
+                            'url': video_url,
+                            'width': int_or_none(video_asset.get('width')),
+                            'height': int_or_none(video_asset.get('height')),
+                            'tbr': tbr,
+                            'container': video_asset.get('format'),
+                        })
+            self._sort_formats(formats)
 
             return {
-                '_type': 'url',
-                # We get the best quality video
-                'url': info['videoAssets'][-1]['publicUrl'],
-                'ie_key': 'ThePlatform',
+                'id': video_id,
+                'title': title,
+                'description': info.get('description'),
+                'thumbnail': info.get('description'),
+                'thumbnail': info.get('thumbnail'),
+                'duration': int_or_none(info.get('duration')),
+                'timestamp': parse_iso8601(info.get('pubDate')),
+                'formats': formats,
+                'subtitles': subtitles,
             }