[youtube] fix hd720 format position

[youtube-dl] / youtube_dl / extractor / tvplay.py

diff --git a/youtube_dl/extractor/tvplay.py b/youtube_dl/extractor/tvplay.py
index c0fec25949c25363e9911af160704928966dd2cb..e09b5f804d897954f4488344d27beaa8a7a2eea6 100644 (file)
--- a/youtube_dl/extractor/tvplay.py
+++ b/youtube_dl/extractor/tvplay.py
@@ -15,7 +15,9 @@ from ..utils import (
     int_or_none,
     parse_iso8601,
     qualities,
+    smuggle_url,
     try_get,
+    unsmuggle_url,
     update_url_query,
 )
 
@@ -224,8 +226,17 @@ class TVPlayIE(InfoExtractor):
     ]
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
+        url, smuggled_data = unsmuggle_url(url, {})
+        self._initialize_geo_bypass({
+            'countries': smuggled_data.get('geo_countries'),
+        })
 
+        video_id = self._match_id(url)
+        geo_country = self._search_regex(
+            r'https?://[^/]+\.([a-z]{2})', url,
+            'geo country', default=None)
+        if geo_country:
+            self._initialize_geo_bypass({'countries': [geo_country.upper()]})
         video = self._download_json(
             'http://playapi.mtgx.tv/v3/videos/%s' % video_id, video_id, 'Downloading video JSON')
 
@@ -264,6 +275,8 @@ class TVPlayIE(InfoExtractor):
                     'ext': ext,
                 }
                 if video_url.startswith('rtmp'):
+                    if smuggled_data.get('skip_rtmp'):
+                        continue
                     m = re.search(
                         r'^(?P<url>rtmp://[^/]+/(?P<app>[^/]+))/(?P<playpath>.+)$', video_url)
                     if not m:
@@ -367,6 +380,10 @@ class ViafreeIE(InfoExtractor):
             'skip_download': True,
         },
         'add_ie': [TVPlayIE.ie_key()],
+    }, {
+        # Different og:image URL schema
+        'url': 'http://www.viafree.se/program/reality/sommaren-med-youtube-stjarnorna/sasong-1/avsnitt-2',
+        'only_matching': True,
     }, {
         'url': 'http://www.viafree.no/programmer/underholdning/det-beste-vorspielet/sesong-2/episode-1',
         'only_matching': True,
@@ -384,17 +401,47 @@ class ViafreeIE(InfoExtractor):
 
         webpage = self._download_webpage(url, video_id)
 
+        data = self._parse_json(
+            self._search_regex(
+                r'(?s)window\.App\s*=\s*({.+?})\s*;\s*</script',
+                webpage, 'data', default='{}'),
+            video_id, transform_source=lambda x: re.sub(
+                r'(?s)function\s+[a-zA-Z_][\da-zA-Z_]*\s*\([^)]*\)\s*{[^}]*}\s*',
+                'null', x), fatal=False)
+
         video_id = None
 
-        thumbnail = self._og_search_thumbnail(webpage, default=None)
-        if thumbnail:
-            video_id = self._search_regex(
-                r'https?://[^/]+/imagecache/(?:[^/]+/)+seasons/\d+/(\d{6,})/',
-                thumbnail, 'video id', default=None)
+        if data:
+            video_id = try_get(
+                data, lambda x: x['context']['dispatcher']['stores'][
+                    'ContentPageProgramStore']['currentVideo']['id'],
+                compat_str)
+
+        # Fallback #1 (extract from og:image URL schema)
+        if not video_id:
+            thumbnail = self._og_search_thumbnail(webpage, default=None)
+            if thumbnail:
+                video_id = self._search_regex(
+                    # Patterns seen:
+                    #  http://cdn.playapi.mtgx.tv/imagecache/600x315/cloud/content-images/inbox/765166/a2e95e5f1d735bab9f309fa345cc3f25.jpg
+                    #  http://cdn.playapi.mtgx.tv/imagecache/600x315/cloud/content-images/seasons/15204/758770/4a5ba509ca8bc043e1ebd1a76131cdf2.jpg
+                    r'https?://[^/]+/imagecache/(?:[^/]+/)+(\d{6,})/',
+                    thumbnail, 'video id', default=None)
 
+        # Fallback #2. Extract from raw JSON string.
+        # May extract wrong video id if relatedClips is present.
         if not video_id:
             video_id = self._search_regex(
                 r'currentVideo["\']\s*:\s*.+?["\']id["\']\s*:\s*["\'](\d{6,})',
                 webpage, 'video id')
 
-        return self.url_result('mtg:%s' % video_id, TVPlayIE.ie_key())
+        return self.url_result(
+            smuggle_url(
+                'mtg:%s' % video_id,
+                {
+                    'geo_countries': [
+                        compat_urlparse.urlparse(url).netloc.rsplit('.', 1)[-1]],
+                    # rtmp host mtgfs.fplive.net for viafree is unresolvable
+                    'skip_rtmp': True,
+                }),
+            ie=TVPlayIE.ie_key(), video_id=video_id)