X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/b238854ce845f3796daac74edab2e8a373e8ba1a..d967b1a589a4ead1c13f0c9e17c77cb9a906e556:/test/test_subtitles.py diff --git a/test/test_subtitles.py b/test/test_subtitles.py index 23a6531..27e763e 100644 --- a/test/test_subtitles.py +++ b/test/test_subtitles.py @@ -1,4 +1,5 @@ #!/usr/bin/env python +from __future__ import unicode_literals # Allow direct execution import os @@ -13,60 +14,67 @@ from youtube_dl.extractor import ( YoutubeIE, DailymotionIE, TEDIE, + VimeoIE, + WallaIE, + CeskaTelevizeIE, + LyndaIE, + NPOIE, + ComedyCentralIE, + NRKTVIE, + RaiTVIE, + VikiIE, + ThePlatformIE, + ThePlatformFeedIE, + RTVEALaCartaIE, + FunnyOrDieIE, + DemocracynowIE, ) class BaseTestSubtitles(unittest.TestCase): url = None IE = None + def setUp(self): self.DL = FakeYDL() - self.ie = self.IE(self.DL) + self.ie = self.IE() + self.DL.add_info_extractor(self.ie) def getInfoDict(self): - info_dict = self.ie.extract(self.url) + info_dict = self.DL.extract_info(self.url, download=False) return info_dict def getSubtitles(self): info_dict = self.getInfoDict() - return info_dict['subtitles'] + subtitles = info_dict['requested_subtitles'] + if not subtitles: + return subtitles + for sub_info in subtitles.values(): + if sub_info.get('data') is None: + uf = self.DL.urlopen(sub_info['url']) + sub_info['data'] = uf.read().decode('utf-8') + return dict((l, sub_info['data']) for l, sub_info in subtitles.items()) class TestYoutubeSubtitles(BaseTestSubtitles): url = 'QRS8MkLhQmM' IE = YoutubeIE - def getSubtitles(self): - info_dict = self.getInfoDict() - return info_dict[0]['subtitles'] - - def test_youtube_no_writesubtitles(self): - self.DL.params['writesubtitles'] = False - subtitles = self.getSubtitles() - self.assertEqual(subtitles, None) - - def test_youtube_subtitles(self): - self.DL.params['writesubtitles'] = True - subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['en']), '4cd9278a35ba2305f47354ee13472260') - - def test_youtube_subtitles_lang(self): - self.DL.params['writesubtitles'] = True - self.DL.params['subtitleslangs'] = ['it'] - subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['it']), '164a51f16f260476a05b50fe4c2f161d') - def test_youtube_allsubtitles(self): self.DL.params['writesubtitles'] = True self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() self.assertEqual(len(subtitles.keys()), 13) + self.assertEqual(md5(subtitles['en']), '3cb210999d3e021bd6c7f0ea751eab06') + self.assertEqual(md5(subtitles['it']), '6d752b98c31f1cf8d597050c7a2cb4b5') + for lang in ['fr', 'de']: + self.assertTrue(subtitles.get(lang) is not None, 'Subtitles for \'%s\' not extracted' % lang) - def test_youtube_subtitles_sbv_format(self): + def test_youtube_subtitles_ttml_format(self): self.DL.params['writesubtitles'] = True - self.DL.params['subtitlesformat'] = 'sbv' + self.DL.params['subtitlesformat'] = 'ttml' subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['en']), '13aeaa0c245a8bed9a451cb643e3ad8b') + self.assertEqual(md5(subtitles['en']), 'e306f8c42842f723447d9f63ad65df54') def test_youtube_subtitles_vtt_format(self): self.DL.params['writesubtitles'] = True @@ -74,12 +82,6 @@ class TestYoutubeSubtitles(BaseTestSubtitles): subtitles = self.getSubtitles() self.assertEqual(md5(subtitles['en']), '3cb210999d3e021bd6c7f0ea751eab06') - def test_youtube_list_subtitles(self): - self.DL.expect_warning(u'Video doesn\'t have automatic captions') - self.DL.params['listsubtitles'] = True - info_dict = self.getInfoDict() - self.assertEqual(info_dict, None) - def test_youtube_automatic_captions(self): self.url = '8YoUxe5ncPo' self.DL.params['writeautomaticsub'] = True @@ -87,124 +89,270 @@ class TestYoutubeSubtitles(BaseTestSubtitles): subtitles = self.getSubtitles() self.assertTrue(subtitles['it'] is not None) - def test_youtube_nosubtitles(self): - self.DL.expect_warning(u'video doesn\'t have subtitles') - self.url = 'sAjKT8FhjI8' - self.DL.params['writesubtitles'] = True - self.DL.params['allsubtitles'] = True + def test_youtube_translated_subtitles(self): + # This video has a subtitles track, which can be translated + self.url = 'Ky9eprVWzlI' + self.DL.params['writeautomaticsub'] = True + self.DL.params['subtitleslangs'] = ['it'] subtitles = self.getSubtitles() - self.assertEqual(len(subtitles), 0) + self.assertTrue(subtitles['it'] is not None) - def test_youtube_multiple_langs(self): - self.url = 'QRS8MkLhQmM' + def test_youtube_nosubtitles(self): + self.DL.expect_warning('video doesn\'t have subtitles') + self.url = 'n5BB19UTcdA' self.DL.params['writesubtitles'] = True - langs = ['it', 'fr', 'de'] - self.DL.params['subtitleslangs'] = langs + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - for lang in langs: - self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang) + self.assertFalse(subtitles) class TestDailymotionSubtitles(BaseTestSubtitles): url = 'http://www.dailymotion.com/video/xczg00' IE = DailymotionIE - def test_no_writesubtitles(self): + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(subtitles, None) + self.assertTrue(len(subtitles.keys()) >= 6) + self.assertEqual(md5(subtitles['en']), '976553874490cba125086bbfea3ff76f') + self.assertEqual(md5(subtitles['fr']), '594564ec7d588942e384e920e5341792') + for lang in ['es', 'fr', 'de']: + self.assertTrue(subtitles.get(lang) is not None, 'Subtitles for \'%s\' not extracted' % lang) - def test_subtitles(self): + def test_nosubtitles(self): + self.DL.expect_warning('video doesn\'t have subtitles') + self.url = 'http://www.dailymotion.com/video/x12u166_le-zapping-tele-star-du-08-aout-2013_tv' self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['en']), '976553874490cba125086bbfea3ff76f') + self.assertFalse(subtitles) - def test_subtitles_lang(self): + +class TestTedSubtitles(BaseTestSubtitles): + url = 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html' + IE = TEDIE + + def test_allsubtitles(self): self.DL.params['writesubtitles'] = True - self.DL.params['subtitleslangs'] = ['fr'] + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['fr']), '594564ec7d588942e384e920e5341792') + self.assertTrue(len(subtitles.keys()) >= 28) + self.assertEqual(md5(subtitles['en']), '4262c1665ff928a2dada178f62cb8d14') + self.assertEqual(md5(subtitles['fr']), '66a63f7f42c97a50f8c0e90bc7797bb5') + for lang in ['es', 'fr', 'de']: + self.assertTrue(subtitles.get(lang) is not None, 'Subtitles for \'%s\' not extracted' % lang) + + +class TestVimeoSubtitles(BaseTestSubtitles): + url = 'http://vimeo.com/76979871' + IE = VimeoIE def test_allsubtitles(self): self.DL.params['writesubtitles'] = True self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(len(subtitles.keys()), 5) + self.assertEqual(set(subtitles.keys()), set(['de', 'en', 'es', 'fr'])) + self.assertEqual(md5(subtitles['en']), '8062383cf4dec168fc40a088aa6d5888') + self.assertEqual(md5(subtitles['fr']), 'b6191146a6c5d3a452244d853fde6dc8') - def test_list_subtitles(self): - self.DL.expect_warning(u'Automatic Captions not supported by this server') - self.DL.params['listsubtitles'] = True - info_dict = self.getInfoDict() - self.assertEqual(info_dict, None) + def test_nosubtitles(self): + self.DL.expect_warning('video doesn\'t have subtitles') + self.url = 'http://vimeo.com/56015672' + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertFalse(subtitles) - def test_automatic_captions(self): - self.DL.expect_warning(u'Automatic Captions not supported by this server') - self.DL.params['writeautomaticsub'] = True - self.DL.params['subtitleslang'] = ['en'] + +class TestWallaSubtitles(BaseTestSubtitles): + url = 'http://vod.walla.co.il/movie/2705958/the-yes-men' + IE = WallaIE + + def test_allsubtitles(self): + self.DL.expect_warning('Automatic Captions not supported by this server') + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertTrue(len(subtitles.keys()) == 0) + self.assertEqual(set(subtitles.keys()), set(['heb'])) + self.assertEqual(md5(subtitles['heb']), 'e758c5d7cb982f6bef14f377ec7a3920') def test_nosubtitles(self): - self.DL.expect_warning(u'video doesn\'t have subtitles') - self.url = 'http://www.dailymotion.com/video/x12u166_le-zapping-tele-star-du-08-aout-2013_tv' + self.DL.expect_warning('video doesn\'t have subtitles') + self.url = 'http://vod.walla.co.il/movie/2642630/one-direction-all-for-one' self.DL.params['writesubtitles'] = True self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(len(subtitles), 0) + self.assertFalse(subtitles) + - def test_multiple_langs(self): +class TestCeskaTelevizeSubtitles(BaseTestSubtitles): + url = 'http://www.ceskatelevize.cz/ivysilani/10600540290-u6-uzasny-svet-techniky' + IE = CeskaTelevizeIE + + def test_allsubtitles(self): + self.DL.expect_warning('Automatic Captions not supported by this server') self.DL.params['writesubtitles'] = True - langs = ['es', 'fr', 'de'] - self.DL.params['subtitleslangs'] = langs + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - for lang in langs: - self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang) + self.assertEqual(set(subtitles.keys()), set(['cs'])) + self.assertTrue(len(subtitles['cs']) > 20000) + def test_nosubtitles(self): + self.DL.expect_warning('video doesn\'t have subtitles') + self.url = 'http://www.ceskatelevize.cz/ivysilani/ivysilani/10441294653-hyde-park-civilizace/214411058091220' + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertFalse(subtitles) -class TestTedSubtitles(BaseTestSubtitles): - url = 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html' - IE = TEDIE - def test_no_writesubtitles(self): +class TestLyndaSubtitles(BaseTestSubtitles): + url = 'http://www.lynda.com/Bootstrap-tutorials/Using-exercise-files/110885/114408-4.html' + IE = LyndaIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(subtitles, None) + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), '09bbe67222259bed60deaa26997d73a7') + - def test_subtitles(self): +class TestNPOSubtitles(BaseTestSubtitles): + url = 'http://www.npo.nl/nos-journaal/28-08-2014/POW_00722860' + IE = NPOIE + + def test_allsubtitles(self): self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['en']), '2154f31ff9b9f89a0aa671537559c21d') + self.assertEqual(set(subtitles.keys()), set(['nl'])) + self.assertEqual(md5(subtitles['nl']), 'fc6435027572b63fb4ab143abd5ad3f4') + - def test_subtitles_lang(self): +class TestMTVSubtitles(BaseTestSubtitles): + url = 'http://www.cc.com/video-clips/kllhuv/stand-up-greg-fitzsimmons--uncensored---too-good-of-a-mother' + IE = ComedyCentralIE + + def getInfoDict(self): + return super(TestMTVSubtitles, self).getInfoDict()['entries'][0] + + def test_allsubtitles(self): self.DL.params['writesubtitles'] = True - self.DL.params['subtitleslangs'] = ['fr'] + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(md5(subtitles['fr']), '7616cbc6df20ec2c1204083c83871cf6') + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), 'b9f6ca22a6acf597ec76f61749765e65') + + +class TestNRKSubtitles(BaseTestSubtitles): + url = 'http://tv.nrk.no/serie/ikke-gjoer-dette-hjemme/DMPV73000411/sesong-2/episode-1' + IE = NRKTVIE def test_allsubtitles(self): self.DL.params['writesubtitles'] = True self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertEqual(len(subtitles.keys()), 28) + self.assertEqual(set(subtitles.keys()), set(['no'])) + self.assertEqual(md5(subtitles['no']), '544fa917d3197fcbee64634559221cc2') - def test_list_subtitles(self): - self.DL.expect_warning(u'Automatic Captions not supported by this server') - self.DL.params['listsubtitles'] = True - info_dict = self.getInfoDict() - self.assertEqual(info_dict, None) - def test_automatic_captions(self): - self.DL.expect_warning(u'Automatic Captions not supported by this server') - self.DL.params['writeautomaticsub'] = True - self.DL.params['subtitleslang'] = ['en'] +class TestRaiSubtitles(BaseTestSubtitles): + url = 'http://www.rai.tv/dl/RaiTV/programmi/media/ContentItem-cb27157f-9dd0-4aee-b788-b1f67643a391.html' + IE = RaiTVIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['it'])) + self.assertEqual(md5(subtitles['it']), 'b1d90a98755126b61e667567a1f6680a') + + +class TestVikiSubtitles(BaseTestSubtitles): + url = 'http://www.viki.com/videos/1060846v-punch-episode-18' + IE = VikiIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), '53cb083a5914b2d84ef1ab67b880d18a') + + +class TestThePlatformSubtitles(BaseTestSubtitles): + # from http://www.3playmedia.com/services-features/tools/integrations/theplatform/ + # (see http://theplatform.com/about/partners/type/subtitles-closed-captioning/) + url = 'theplatform:JFUjUE1_ehvq' + IE = ThePlatformIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - self.assertTrue(len(subtitles.keys()) == 0) + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), '97e7670cbae3c4d26ae8bcc7fdd78d4b') + + +class TestThePlatformFeedSubtitles(BaseTestSubtitles): + url = 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207' + IE = ThePlatformFeedIE - def test_multiple_langs(self): + def test_allsubtitles(self): self.DL.params['writesubtitles'] = True - langs = ['es', 'fr', 'de'] - self.DL.params['subtitleslangs'] = langs + self.DL.params['allsubtitles'] = True subtitles = self.getSubtitles() - for lang in langs: - self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang) + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), '48649a22e82b2da21c9a67a395eedade') + + +class TestRtveSubtitles(BaseTestSubtitles): + url = 'http://www.rtve.es/alacarta/videos/los-misterios-de-laura/misterios-laura-capitulo-32-misterio-del-numero-17-2-parte/2428621/' + IE = RTVEALaCartaIE + + def test_allsubtitles(self): + print('Skipping, only available from Spain') + return + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['es'])) + self.assertEqual(md5(subtitles['es']), '69e70cae2d40574fb7316f31d6eb7fca') + + +class TestFunnyOrDieSubtitles(BaseTestSubtitles): + url = 'http://www.funnyordie.com/videos/224829ff6d/judd-apatow-will-direct-your-vine' + IE = FunnyOrDieIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), 'c5593c193eacd353596c11c2d4f9ecc4') + + +class TestDemocracynowSubtitles(BaseTestSubtitles): + url = 'http://www.democracynow.org/shows/2015/7/3' + IE = DemocracynowIE + + def test_allsubtitles(self): + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), 'acaca989e24a9e45a6719c9b3d60815c') + + def test_subtitles_in_page(self): + self.url = 'http://www.democracynow.org/2015/7/3/this_flag_comes_down_today_bree' + self.DL.params['writesubtitles'] = True + self.DL.params['allsubtitles'] = True + subtitles = self.getSubtitles() + self.assertEqual(set(subtitles.keys()), set(['en'])) + self.assertEqual(md5(subtitles['en']), 'acaca989e24a9e45a6719c9b3d60815c') + if __name__ == '__main__': unittest.main()