switched ytsearch to more robust Youtube Data API (fixes #307)
authorFilippo Valsorda <filippo.valsorda@gmail.com>
Wed, 14 Mar 2012 21:44:45 +0000 (22:44 +0100)
committerFilippo Valsorda <filippo.valsorda@gmail.com>
Wed, 14 Mar 2012 21:44:45 +0000 (22:44 +0100)
youtube-dl

index 1b381d7b75307008b3686e036417edf155d896e7..a1871ca1c2a17920de98ca0e13c4f188fa396c84 100755 (executable)
@@ -2248,9 +2248,7 @@ class GenericIE(InfoExtractor):
 class YoutubeSearchIE(InfoExtractor):
        """Information Extractor for YouTube search queries."""
        _VALID_URL = r'ytsearch(\d+|all)?:[\s\S]+'
-       _TEMPLATE_URL = 'http://www.youtube.com/results?search_query=%s&page=%s&gl=US&hl=en'
-       _VIDEO_INDICATOR = r'href="/watch\?v=.+?"'
-       _MORE_PAGES_INDICATOR = r'(?m)>\s*Next\s*</a>'
+       _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
        _youtube_ie = None
        _max_youtube_results = 1000
        IE_NAME = u'youtube:search'
@@ -2301,37 +2299,31 @@ class YoutubeSearchIE(InfoExtractor):
                """Downloads a specified number of results for a query"""
 
                video_ids = []
-               already_seen = set()
-               pagenum = 1
+               pagenum = 0
+               limit = n
 
-               while True:
-                       self.report_download_page(query, pagenum)
-                       result_url = self._TEMPLATE_URL % (urllib.quote_plus(query), pagenum)
+               while (50 * pagenum) < limit:
+                       self.report_download_page(query, pagenum+1)
+                       result_url = self._API_URL % (urllib.quote_plus(query), (50*pagenum)+1)
                        request = urllib2.Request(result_url)
                        try:
-                               page = urllib2.urlopen(request).read()
+                               data = urllib2.urlopen(request).read()
                        except (urllib2.URLError, httplib.HTTPException, socket.error), err:
-                               self._downloader.trouble(u'ERROR: unable to download webpage: %s' % str(err))
+                               self._downloader.trouble(u'ERROR: unable to download API page: %s' % str(err))
                                return
+                       api_response = json.loads(data)['data']
 
-                       # Extract video identifiers
-                       for mobj in re.finditer(self._VIDEO_INDICATOR, page):
-                               video_id = page[mobj.span()[0]:mobj.span()[1]].split('=')[2][:-1]
-                               if video_id not in already_seen:
-                                       video_ids.append(video_id)
-                                       already_seen.add(video_id)
-                                       if len(video_ids) == n:
-                                               # Specified n videos reached
-                                               for id in video_ids:
-                                                       self._youtube_ie.extract('http://www.youtube.com/watch?v=%s' % id)
-                                               return
+                       new_ids = list(video['id'] for video in api_response['items'])
+                       video_ids += new_ids
 
-                       if re.search(self._MORE_PAGES_INDICATOR, page) is None:
-                               for id in video_ids:
-                                       self._youtube_ie.extract('http://www.youtube.com/watch?v=%s' % id)
-                               return
+                       limit = min(n, api_response['totalItems'])
+                       pagenum += 1
 
-                       pagenum = pagenum + 1
+               if len(video_ids) > n:
+                       video_ids = video_ids[:n]
+               for id in video_ids:
+                       self._youtube_ie.extract('http://www.youtube.com/watch?v=%s' % id)
+               return
 
 
 class GoogleSearchIE(InfoExtractor):