[screenjunkies] Improve, extract more metadata and workaround subscription (Closes...
authorSergey M․ <dstftw@gmail.com>
Mon, 15 Feb 2016 16:29:28 +0000 (22:29 +0600)
committerSergey M․ <dstftw@gmail.com>
Mon, 15 Feb 2016 16:29:28 +0000 (22:29 +0600)
youtube_dl/extractor/screenjunkies.py

index 5da3f9bcde14c827e3293fbb2da6e15ecae0c678..e0db180b81992bea04380a1e704e32e1ae70575c 100644 (file)
 from __future__ import unicode_literals
 
 import re
-import json
-
-from .common import (
-    InfoExtractor,
-    ExtractorError,
-)
 
+from .common import InfoExtractor
+from ..compat import compat_str
 from ..utils import (
     int_or_none,
+    parse_age_limit,
 )
 
-class ScreenJunkiesIE(InfoExtractor):
-    _VALID_URL = r'http://www.screenjunkies.com/video/(.+-(?P<id>\d+)|.+)'
 
+class ScreenJunkiesIE(InfoExtractor):
+    _VALID_URL = r'http://www.screenjunkies.com/video/(?P<display_id>[^/]+?)(?:-(?P<id>\d+))?(?:[/?#&]|$)'
     _TESTS = [{
         'url': 'http://www.screenjunkies.com/video/best-quentin-tarantino-movie-2841915',
+        'md5': '5c2b686bec3d43de42bde9ec047536b0',
         'info_dict': {
             'id': '2841915',
+            'display_id': 'best-quentin-tarantino-movie',
             'ext': 'mp4',
             'title': 'Best Quentin Tarantino Movie',
+            'thumbnail': 're:^https?://.*\.jpg',
+            'duration': 3671,
+            'age_limit': 13,
+            'tags': list,
         },
     }, {
         'url': 'http://www.screenjunkies.com/video/honest-trailers-the-dark-knight',
         'info_dict': {
             'id': '2348808',
+            'display_id': 'honest-trailers-the-dark-knight',
             'ext': 'mp4',
-            'title': "Honest Trailers - 'The Dark Knight'",
+            'title': "Honest Trailers: 'The Dark Knight'",
+            'thumbnail': 're:^https?://.*\.jpg',
+            'age_limit': 10,
+            'tags': list,
         },
     }, {
+        # requires subscription but worked around
         'url': 'http://www.screenjunkies.com/video/knocking-dead-ep-1-the-show-so-far-3003285',
-        'only_matching': True,
+        'info_dict': {
+            'id': '3003285',
+            'display_id': 'knocking-dead-ep-1-the-show-so-far',
+            'ext': 'mp4',
+            'title': 'Knocking Dead Ep 1: State of The Dead Recap',
+            'thumbnail': 're:^https?://.*\.jpg',
+            'duration': 3307,
+            'age_limit': 13,
+            'tags': list,
+        },
     }]
 
+    _DEFAULT_BITRATES = (48, 150, 496, 864, 2240)
+
     def _real_extract(self, url):
-        video_id = self._match_id(url)
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        display_id = mobj.group('display_id')
 
-        if not video_id: # Older urls didn't have the id in them, but we can grab it manually
-            webpage = self._download_webpage(url, url)
-            video_id = self._html_search_regex(r'src="/embed/(\d+)"', webpage, 'video id')
+        if not video_id:
+            webpage = self._download_webpage(url, display_id)
+            video_id = self._search_regex(
+                (r'src=["\']/embed/(\d+)', r'data-video-content-id=["\'](\d+)'),
+                webpage, 'video id')
 
-        webpage = self._download_webpage('http://www.screenjunkies.com/embed/%s' %video_id, video_id)
-        video_vars_str = self._html_search_regex(r'embedVars = (\{.+\})\s*</script>', webpage, 'video variables', flags=re.S)
-        video_vars = self._parse_json(video_vars_str, video_id)
+        webpage = self._download_webpage(
+            'http://www.screenjunkies.com/embed/%s' % video_id,
+            display_id, 'Downloading video embed page')
+        embed_vars = self._parse_json(
+            self._search_regex(
+                r'(?s)embedVars\s*=\s*({.+?})\s*</script>', webpage, 'embed vars'),
+            display_id)
 
-        # TODO: Figure out required cookies
-        if video_vars['subscriptionLevel'] > 0:
-            raise ExtractorError('This video requires ScreenJunkiesPlus', expected=True)
+        title = embed_vars['contentName']
 
         formats = []
-        for f in video_vars['media']:
-            if not f['mediaPurpose'] == 'play':
+        bitrates = []
+        for f in embed_vars.get('media', []):
+            if not f.get('uri') or f.get('mediaPurpose') != 'play':
                 continue
-
+            bitrate = int_or_none(f.get('bitRate'))
+            if bitrate:
+                bitrates.append(bitrate)
             formats.append({
                 'url': f['uri'],
+                'format_id': 'http-%d' % bitrate if bitrate else 'http',
                 'width': int_or_none(f.get('width')),
                 'height': int_or_none(f.get('height')),
-                'tbr': int_or_none(f.get('bitRate')),
+                'tbr': bitrate,
                 'format': 'mp4',
             })
+
+        if not bitrates:
+            # When subscriptionLevel > 0, i.e. plus subscription is required
+            # media list will be empty. However, hds and hls uri are still
+            # available. We can grabbing them assuming bitrates to be default.
+            bitrates = self._DEFAULT_BITRATES
+
+        auth_token = embed_vars.get('AuthToken')
+
+        def construct_manifest_url(base_url, ext):
+            pieces = [base_url]
+            pieces.extend([compat_str(b) for b in bitrates])
+            pieces.append('_kbps.mp4.%s?%s' % (ext, auth_token))
+            return ','.join(pieces)
+
+        if bitrates and auth_token:
+            hds_url = embed_vars.get('hdsUri')
+            if hds_url:
+                f4m_formats = self._extract_f4m_formats(
+                    construct_manifest_url(hds_url, 'f4m'),
+                    display_id, f4m_id='hds', fatal=False)
+                if len(f4m_formats) == len(bitrates):
+                    for f, bitrate in zip(f4m_formats, bitrates):
+                        if not f.get('tbr'):
+                            f['format_id'] = 'hds-%d' % bitrate
+                            f['tbr'] = bitrate
+                # TODO: fix f4m downloader to handle manifests without bitrates if possible
+                # formats.extend(f4m_formats)
+
+            hls_url = embed_vars.get('hlsUri')
+            if hls_url:
+                formats.extend(self._extract_m3u8_formats(
+                    construct_manifest_url(hls_url, 'm3u8'),
+                    display_id, 'mp4', entry_protocol='m3u8_native', m3u8_id='hls', fatal=False))
         self._sort_formats(formats)
 
         return {
             'id': video_id,
-            'title': video_vars['contentName'],
+            'display_id': display_id,
+            'title': title,
+            'thumbnail': embed_vars.get('thumbUri'),
+            'duration': int_or_none(embed_vars.get('videoLengthInSeconds')) or None,
+            'age_limit': parse_age_limit(embed_vars.get('audienceRating')),
+            'tags': embed_vars.get('tags', '').split(','),
             'formats': formats,
-            'duration': int_or_none(video_vars.get('videoLengthInSeconds')),
-            'thumbnail': video_vars.get('thumbUri'),
-            'tags': video_vars['tags'].split(',') if 'tags' in video_vars else [],
         }
-