From a3978a615950af6e990313820f93baddce067ee4 Mon Sep 17 00:00:00 2001 From: dst Date: Tue, 28 Jan 2014 09:12:23 +0700 Subject: [PATCH 1/2] [imdb] Fix duplicated entries bug --- youtube_dl/extractor/imdb.py | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/youtube_dl/extractor/imdb.py b/youtube_dl/extractor/imdb.py index 1763af020..7cee505c0 100644 --- a/youtube_dl/extractor/imdb.py +++ b/youtube_dl/extractor/imdb.py @@ -69,12 +69,9 @@ class ImdbListIE(InfoExtractor): list_id = mobj.group('id') webpage = self._download_webpage(url, list_id) - list_code = self._search_regex( - r'(?s)(.*?)class="see-more"', - webpage, 'list code') entries = [ self.url_result('http://www.imdb.com' + m, 'Imdb') - for m in re.findall(r'href="(/video/imdb/vi[^"]+)"', webpage)] + for m in re.findall(r'href="(/video/imdb/vi[^"]+)"\s+data-type="playlist"', webpage)] list_title = self._html_search_regex( r'

(.*?)

', webpage, 'list title') From 98669ed79c10bbcfc5678a520de98df79ca6b1b4 Mon Sep 17 00:00:00 2001 From: dst Date: Tue, 28 Jan 2014 09:13:08 +0700 Subject: [PATCH 2/2] [imdb] Fix playlist test --- test/test_playlists.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/test/test_playlists.py b/test/test_playlists.py index 5eeba091e..a8f4a49f4 100644 --- a/test/test_playlists.py +++ b/test/test_playlists.py @@ -195,11 +195,11 @@ class TestPlaylists(unittest.TestCase): def test_imdb_list(self): dl = FakeYDL() ie = ImdbListIE(dl) - result = ie.extract('http://www.imdb.com/list/sMjedvGDd8U') + result = ie.extract('http://www.imdb.com/list/JFs9NWw6XI0') self.assertIsPlaylist(result) - self.assertEqual(result['id'], 'sMjedvGDd8U') - self.assertEqual(result['title'], 'Animated and Family Films') - self.assertTrue(len(result['entries']) >= 48) + self.assertEqual(result['id'], 'JFs9NWw6XI0') + self.assertEqual(result['title'], 'March 23, 2012 Releases') + self.assertEqual(len(result['entries']), 7) def test_khanacademy_topic(self): dl = FakeYDL()