[yahoo] Add an extractor for yahoo news (closes #1849)

[youtube-dl.git] / youtube_dl / extractor / yahoo.py

diff --git a/youtube_dl/extractor/yahoo.py b/youtube_dl/extractor/yahoo.py
index 5bdd5d591bbdbabd4abb28ca0625de70c4ecb6c9..2d87e81b2d6072a80129062bfe8142c19bf40aa5 100644 (file)
--- a/youtube_dl/extractor/yahoo.py
+++ b/youtube_dl/extractor/yahoo.py
@@ -46,10 +46,28 @@ class YahooIE(InfoExtractor):
         video_id = mobj.group('id')
         webpage = self._download_webpage(url, video_id)
 
-        items_json = self._search_regex(r'YVIDEO_INIT_ITEMS = ({.*?});$',
+        items_json = self._search_regex(r'mediaItems: ({.*?})$',
             webpage, u'items', flags=re.MULTILINE)
         items = json.loads(items_json)
         info = items['mediaItems']['query']['results']['mediaObj'][0]
+        # The 'meta' field is not always in the video webpage, we request it
+        # from another page
+        long_id = info['id']
+        return self._get_info(info['id'], video_id)
+
+    def _get_info(self, long_id, video_id):
+        query = ('SELECT * FROM yahoo.media.video.streams WHERE id="%s"'
+                 ' AND plrs="86Gj0vCaSzV_Iuf6hNylf2" AND region="US"' % long_id)
+        data = compat_urllib_parse.urlencode({
+            'q': query,
+            'env': 'prod',
+            'format': 'json',
+        })
+        query_result_json = self._download_webpage(
+            'http://video.query.yahoo.com/v1/public/yql?' + data,
+            video_id, u'Downloading video info')
+        query_result = json.loads(query_result_json)
+        info = query_result['query']['results']['mediaObj'][0]
         meta = info['meta']
 
         formats = []
@@ -76,17 +94,42 @@ class YahooIE(InfoExtractor):
             formats.append(format_info)
         formats = sorted(formats, key=lambda f:(f['height'], f['width']))
 
-        info = {
+        return {
             'id': video_id,
             'title': meta['title'],
             'formats': formats,
             'description': clean_html(meta['description']),
             'thumbnail': meta['thumbnail'],
         }
-        # TODO: Remove when #980 has been merged
-        info.update(formats[-1])
 
-        return info
+
+class YahooNewsIE(YahooIE):
+    IE_NAME = 'yahoo:news'
+    _VALID_URL = r'http://news\.yahoo\.com/video/.*?-(?P<id>\d*?)\.html'
+
+    _TEST = {
+        u'url': u'http://news.yahoo.com/video/china-moses-crazy-blues-104538833.html',
+        u'info_dict': {
+            u'id': u'104538833',
+            u'ext': u'flv',
+            u'title': u'China Moses Is Crazy About the Blues',
+            u'description': u'md5:9900ab8cd5808175c7b3fe55b979bed0',
+        },
+        u'params': {
+            # Requires rtmpdump
+            u'skip_download': True,
+        },
+    }
+
+    # Overwrite YahooIE properties we don't want
+    _TESTS = []
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        webpage = self._download_webpage(url, video_id)
+        long_id = self._search_regex(r'contentId: \'(.+?)\',', webpage, u'long id')
+        return self._get_info(long_id, video_id)
 
 
 class YahooSearchIE(SearchInfoExtractor):
@@ -117,7 +160,7 @@ class YahooSearchIE(SearchInfoExtractor):
                 mobj = re.search(r'(?P<url>screen\.yahoo\.com/.*?-\d*?\.html)"', r)
                 e = self.url_result('http://' + mobj.group('url'), 'Yahoo')
                 res['entries'].append(e)
-            if (pagenum * 30 +i >= n) or (m[u'last'] >= (m[u'total'] -1 )):
+            if (pagenum * 30 +i >= n) or (m[u'last'] >= (m[u'total'] -1)):
                 break
 
         return res