X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/9a117f94b4bfe84cfe1d904d5132aefcf41511c9..67afe88251ff8b1fcb43246c9a4bb1a0ea0185a2:/youtube_dl/extractor/nba.py

diff --git a/youtube_dl/extractor/nba.py b/youtube_dl/extractor/nba.py
index 122b7dd..944096e 100644
--- a/youtube_dl/extractor/nba.py
+++ b/youtube_dl/extractor/nba.py
@@ -1,49 +1,63 @@
-import re
+from __future__ import unicode_literals
 
 from .common import InfoExtractor
 from ..utils import (
-    ExtractorError,
+    remove_end,
+    parse_duration,
 )
 
 
 class NBAIE(InfoExtractor):
-    _VALID_URL = r'^(?:https?://)?(?:watch\.|www\.)?nba\.com/(?:nba/)?video(/[^?]*?)(?:/index\.html)?(?:\?.*)?$'
-    _TEST = {
-        u'url': u'http://www.nba.com/video/games/nets/2012/12/04/0021200253-okc-bkn-recap.nba/index.html',
-        u'file': u'0021200253-okc-bkn-recap.nba.mp4',
-        u'md5': u'c0edcfc37607344e2ff8f13c378c88a4',
-        u'info_dict': {
-            u"description": u"Kevin Durant scores 32 points and dishes out six assists as the Thunder beat the Nets in Brooklyn.", 
-            u"title": u"Thunder vs. Nets"
+    _VALID_URL = r'https?://(?:watch\.|www\.)?nba\.com/(?:nba/)?video(?P<id>/[^?]*?)/?(?:/index\.html)?(?:\?.*)?$'
+    _TESTS = [{
+        'url': 'http://www.nba.com/video/games/nets/2012/12/04/0021200253-okc-bkn-recap.nba/index.html',
+        'md5': 'c0edcfc37607344e2ff8f13c378c88a4',
+        'info_dict': {
+            'id': '0021200253-okc-bkn-recap.nba',
+            'ext': 'mp4',
+            'title': 'Thunder vs. Nets',
+            'description': 'Kevin Durant scores 32 points and dishes out six assists as the Thunder beat the Nets in Brooklyn.',
+            'duration': 181,
+        },
+    }, {
+        'url': 'http://www.nba.com/video/games/hornets/2014/12/05/0021400276-nyk-cha-play5.nba/',
+        'only_matching': True,
+    }, {
+        'url': 'http://watch.nba.com/nba/video/channels/playoffs/2015/05/20/0041400301-cle-atl-recap.nba',
+        'info_dict': {
+            'id': '0041400301-cle-atl-recap.nba',
+            'ext': 'mp4',
+            'title': 'NBA GAME TIME | Video: Hawks vs. Cavaliers Game 1',
+            'description': 'md5:8094c3498d35a9bd6b1a8c396a071b4d',
+            'duration': 228,
+        },
+        'params': {
+            'skip_download': True,
         }
-    }
+    }]
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        if mobj is None:
-            raise ExtractorError(u'Invalid URL: %s' % url)
-
-        video_id = mobj.group(1)
-
+        video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
 
-        video_url = u'http://ht-mobile.cdn.turner.com/nba/big' + video_id + '_nba_1280x720.mp4'
+        video_url = 'http://ht-mobile.cdn.turner.com/nba/big' + video_id + '_nba_1280x720.mp4'
 
         shortened_video_id = video_id.rpartition('/')[2]
-        title = self._html_search_regex(r'<meta property="og:title" content="(.*?)"',
-            webpage, 'title', default=shortened_video_id).replace('NBA.com: ', '')
-
-        # It isn't there in the HTML it returns to us
-        # uploader_date = self._html_search_regex(r'<b>Date:</b> (.*?)</div>', webpage, 'upload_date', fatal=False)
-
-        description = self._html_search_regex(r'<meta name="description" (?:content|value)="(.*?)" />', webpage, 'description', fatal=False)
-
-        info = {
+        title = remove_end(
+            self._og_search_title(webpage, default=shortened_video_id), ' : NBA.com')
+
+        description = self._og_search_description(webpage)
+        duration_str = self._html_search_meta(
+            'duration', webpage, 'duration', default=None)
+        if not duration_str:
+            duration_str = self._html_search_regex(
+                r'Duration:</b>\s*(\d+:\d+)', webpage, 'duration', fatal=False)
+        duration = parse_duration(duration_str)
+
+        return {
             'id': shortened_video_id,
             'url': video_url,
-            'ext': 'mp4',
             'title': title,
-            # 'uploader_date': uploader_date,
             'description': description,
+            'duration': duration,
         }
-        return [info]