[yt-dlp.git] / test / test_subtitles.py

#!/usr/bin/env python

# Allow direct execution
import os
import sys
import unittest
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from test.helper import FakeYDL, md5


from youtube_dl.extractor import (
    BlipTVIE,
    YoutubeIE,
    DailymotionIE,
    TEDIE,
)


class BaseTestSubtitles(unittest.TestCase):
    url = None
    IE = None
    def setUp(self):
        self.DL = FakeYDL()
        self.ie = self.IE(self.DL)

    def getInfoDict(self):
        info_dict = self.ie.extract(self.url)
        return info_dict

    def getSubtitles(self):
        info_dict = self.getInfoDict()
        return info_dict['subtitles']


class TestYoutubeSubtitles(BaseTestSubtitles):
    url = 'QRS8MkLhQmM'
    IE = YoutubeIE

    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)

    def test_youtube_subtitles_sbv_format(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['subtitlesformat'] = 'sbv'
        subtitles = self.getSubtitles()
        self.assertEqual(md5(subtitles['en']), '13aeaa0c245a8bed9a451cb643e3ad8b')

    def test_youtube_subtitles_vtt_format(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['subtitlesformat'] = 'vtt'
        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
        self.DL.params['subtitleslangs'] = ['it']
        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
        subtitles = self.getSubtitles()
        self.assertEqual(len(subtitles), 0)

    def test_youtube_multiple_langs(self):
        self.url = 'QRS8MkLhQmM'
        self.DL.params['writesubtitles'] = True
        langs = ['it', 'fr', 'de']
        self.DL.params['subtitleslangs'] = langs
        subtitles = self.getSubtitles()
        for lang in langs:
            self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)


class TestDailymotionSubtitles(BaseTestSubtitles):
    url = 'http://www.dailymotion.com/video/xczg00'
    IE = DailymotionIE

    def test_no_writesubtitles(self):
        subtitles = self.getSubtitles()
        self.assertEqual(subtitles, None)

    def test_subtitles(self):
        self.DL.params['writesubtitles'] = True
        subtitles = self.getSubtitles()
        self.assertEqual(md5(subtitles['en']), '976553874490cba125086bbfea3ff76f')

    def test_subtitles_lang(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['subtitleslangs'] = ['fr']
        subtitles = self.getSubtitles()
        self.assertEqual(md5(subtitles['fr']), '594564ec7d588942e384e920e5341792')

    def test_allsubtitles(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['allsubtitles'] = True
        subtitles = self.getSubtitles()
        self.assertEqual(len(subtitles.keys()), 5)

    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']
        subtitles = self.getSubtitles()
        self.assertTrue(len(subtitles.keys()) == 0)

    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.params['writesubtitles'] = True
        self.DL.params['allsubtitles'] = True
        subtitles = self.getSubtitles()
        self.assertEqual(len(subtitles), 0)

    def test_multiple_langs(self):
        self.DL.params['writesubtitles'] = True
        langs = ['es', 'fr', 'de']
        self.DL.params['subtitleslangs'] = langs
        subtitles = self.getSubtitles()
        for lang in langs:
            self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)


class TestTedSubtitles(BaseTestSubtitles):
    url = 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html'
    IE = TEDIE

    def test_no_writesubtitles(self):
        subtitles = self.getSubtitles()
        self.assertEqual(subtitles, None)

    def test_subtitles(self):
        self.DL.params['writesubtitles'] = True
        subtitles = self.getSubtitles()
        self.assertEqual(md5(subtitles['en']), '4262c1665ff928a2dada178f62cb8d14')

    def test_subtitles_lang(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['subtitleslangs'] = ['fr']
        subtitles = self.getSubtitles()
        self.assertEqual(md5(subtitles['fr']), '66a63f7f42c97a50f8c0e90bc7797bb5')

    def test_allsubtitles(self):
        self.DL.params['writesubtitles'] = True
        self.DL.params['allsubtitles'] = True
        subtitles = self.getSubtitles()
        self.assertEqual(len(subtitles.keys()), 28)

    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']
        subtitles = self.getSubtitles()
        self.assertTrue(len(subtitles.keys()) == 0)

    def test_multiple_langs(self):
        self.DL.params['writesubtitles'] = True
        langs = ['es', 'fr', 'de']
        self.DL.params['subtitleslangs'] = langs
        subtitles = self.getSubtitles()
        for lang in langs:
            self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)


class TestBlipTVSubtitles(BaseTestSubtitles):
    url = 'http://blip.tv/a/a-6603250'
    IE = BlipTVIE

    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_allsubtitles(self):
        self.DL.expect_warning(u'Automatic Captions not supported by this server')
        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']), '5b75c300af65fe4476dff79478bb93e4')


if __name__ == '__main__':
    unittest.main()
Commit	Line	Data
ae84f879 JMF	1	#!/usr/bin/env python
	2
	3	# Allow direct execution
	4	import os
	5	import sys
	6	import unittest
	7	sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
	8
d0efb9ec	9	from test.helper import FakeYDL, md5
ae84f879 JMF	10
	11
	12	from youtube_dl.extractor import (
b4bcffef	13	BlipTVIE,
ae84f879 JMF	14	YoutubeIE,
	15	DailymotionIE,
	16	TEDIE,
	17	)
	18
	19
	20	class BaseTestSubtitles(unittest.TestCase):
	21	url = None
	22	IE = None
	23	def setUp(self):
	24	self.DL = FakeYDL()
	25	self.ie = self.IE(self.DL)
	26
	27	def getInfoDict(self):
	28	info_dict = self.ie.extract(self.url)
	29	return info_dict
	30
	31	def getSubtitles(self):
	32	info_dict = self.getInfoDict()
	33	return info_dict['subtitles']
	34
	35
	36	class TestYoutubeSubtitles(BaseTestSubtitles):
	37	url = 'QRS8MkLhQmM'
	38	IE = YoutubeIE
	39
ae84f879 JMF	40	def test_youtube_no_writesubtitles(self):
	41	self.DL.params['writesubtitles'] = False
	42	subtitles = self.getSubtitles()
	43	self.assertEqual(subtitles, None)
	44
	45	def test_youtube_subtitles(self):
	46	self.DL.params['writesubtitles'] = True
	47	subtitles = self.getSubtitles()
	48	self.assertEqual(md5(subtitles['en']), '4cd9278a35ba2305f47354ee13472260')
	49
	50	def test_youtube_subtitles_lang(self):
	51	self.DL.params['writesubtitles'] = True
	52	self.DL.params['subtitleslangs'] = ['it']
	53	subtitles = self.getSubtitles()
	54	self.assertEqual(md5(subtitles['it']), '164a51f16f260476a05b50fe4c2f161d')
	55
	56	def test_youtube_allsubtitles(self):
	57	self.DL.params['writesubtitles'] = True
	58	self.DL.params['allsubtitles'] = True
	59	subtitles = self.getSubtitles()
	60	self.assertEqual(len(subtitles.keys()), 13)
	61
	62	def test_youtube_subtitles_sbv_format(self):
	63	self.DL.params['writesubtitles'] = True
	64	self.DL.params['subtitlesformat'] = 'sbv'
	65	subtitles = self.getSubtitles()
	66	self.assertEqual(md5(subtitles['en']), '13aeaa0c245a8bed9a451cb643e3ad8b')
	67
	68	def test_youtube_subtitles_vtt_format(self):
	69	self.DL.params['writesubtitles'] = True
	70	self.DL.params['subtitlesformat'] = 'vtt'
	71	subtitles = self.getSubtitles()
4e0084d9	72	self.assertEqual(md5(subtitles['en']), '3cb210999d3e021bd6c7f0ea751eab06')
ae84f879 JMF	73
	74	def test_youtube_list_subtitles(self):
	75	self.DL.expect_warning(u'Video doesn\'t have automatic captions')
	76	self.DL.params['listsubtitles'] = True
	77	info_dict = self.getInfoDict()
	78	self.assertEqual(info_dict, None)
	79
	80	def test_youtube_automatic_captions(self):
	81	self.url = '8YoUxe5ncPo'
	82	self.DL.params['writeautomaticsub'] = True
	83	self.DL.params['subtitleslangs'] = ['it']
	84	subtitles = self.getSubtitles()
	85	self.assertTrue(subtitles['it'] is not None)
	86
	87	def test_youtube_nosubtitles(self):
	88	self.DL.expect_warning(u'video doesn\'t have subtitles')
	89	self.url = 'sAjKT8FhjI8'
	90	self.DL.params['writesubtitles'] = True
	91	self.DL.params['allsubtitles'] = True
	92	subtitles = self.getSubtitles()
	93	self.assertEqual(len(subtitles), 0)
	94
	95	def test_youtube_multiple_langs(self):
	96	self.url = 'QRS8MkLhQmM'
	97	self.DL.params['writesubtitles'] = True
	98	langs = ['it', 'fr', 'de']
	99	self.DL.params['subtitleslangs'] = langs
	100	subtitles = self.getSubtitles()
	101	for lang in langs:
	102	self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)
	103
	104
	105	class TestDailymotionSubtitles(BaseTestSubtitles):
	106	url = 'http://www.dailymotion.com/video/xczg00'
	107	IE = DailymotionIE
	108
	109	def test_no_writesubtitles(self):
	110	subtitles = self.getSubtitles()
	111	self.assertEqual(subtitles, None)
	112
	113	def test_subtitles(self):
	114	self.DL.params['writesubtitles'] = True
	115	subtitles = self.getSubtitles()
	116	self.assertEqual(md5(subtitles['en']), '976553874490cba125086bbfea3ff76f')
	117
	118	def test_subtitles_lang(self):
	119	self.DL.params['writesubtitles'] = True
	120	self.DL.params['subtitleslangs'] = ['fr']
	121	subtitles = self.getSubtitles()
	122	self.assertEqual(md5(subtitles['fr']), '594564ec7d588942e384e920e5341792')
	123
	124	def test_allsubtitles(self):
	125	self.DL.params['writesubtitles'] = True
	126	self.DL.params['allsubtitles'] = True
	127	subtitles = self.getSubtitles()
	128	self.assertEqual(len(subtitles.keys()), 5)
	129
	130	def test_list_subtitles(self):
	131	self.DL.expect_warning(u'Automatic Captions not supported by this server')
	132	self.DL.params['listsubtitles'] = True
	133	info_dict = self.getInfoDict()
	134	self.assertEqual(info_dict, None)
	135
	136	def test_automatic_captions(self):
137	self.DL.expect_warning(u'Automatic Captions not supported by this server')
138	self.DL.params['writeautomaticsub'] = True
139	self.DL.params['subtitleslang'] = ['en']
140	subtitles = self.getSubtitles()
141	self.assertTrue(len(subtitles.keys()) == 0)
142
143	def test_nosubtitles(self):
144	self.DL.expect_warning(u'video doesn\'t have subtitles')
145	self.url = 'http://www.dailymotion.com/video/x12u166_le-zapping-tele-star-du-08-aout-2013_tv'
146	self.DL.params['writesubtitles'] = True
147	self.DL.params['allsubtitles'] = True
148	subtitles = self.getSubtitles()
149	self.assertEqual(len(subtitles), 0)
150
151	def test_multiple_langs(self):
152	self.DL.params['writesubtitles'] = True
153	langs = ['es', 'fr', 'de']
154	self.DL.params['subtitleslangs'] = langs
155	subtitles = self.getSubtitles()
156	for lang in langs:
157	self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)
158
159
160	class TestTedSubtitles(BaseTestSubtitles):
161	url = 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html'
162	IE = TEDIE
163
164	def test_no_writesubtitles(self):
165	subtitles = self.getSubtitles()
166	self.assertEqual(subtitles, None)
167
168	def test_subtitles(self):
169	self.DL.params['writesubtitles'] = True
170	subtitles = self.getSubtitles()
26dca166	171	self.assertEqual(md5(subtitles['en']), '4262c1665ff928a2dada178f62cb8d14')
ae84f879 JMF	172
	173	def test_subtitles_lang(self):
	174	self.DL.params['writesubtitles'] = True
	175	self.DL.params['subtitleslangs'] = ['fr']
	176	subtitles = self.getSubtitles()
26dca166	177	self.assertEqual(md5(subtitles['fr']), '66a63f7f42c97a50f8c0e90bc7797bb5')
ae84f879 JMF	178
	179	def test_allsubtitles(self):
	180	self.DL.params['writesubtitles'] = True
	181	self.DL.params['allsubtitles'] = True
	182	subtitles = self.getSubtitles()
	183	self.assertEqual(len(subtitles.keys()), 28)
	184
	185	def test_list_subtitles(self):
	186	self.DL.expect_warning(u'Automatic Captions not supported by this server')
	187	self.DL.params['listsubtitles'] = True
	188	info_dict = self.getInfoDict()
	189	self.assertEqual(info_dict, None)
	190
	191	def test_automatic_captions(self):
	192	self.DL.expect_warning(u'Automatic Captions not supported by this server')
	193	self.DL.params['writeautomaticsub'] = True
	194	self.DL.params['subtitleslang'] = ['en']
	195	subtitles = self.getSubtitles()
	196	self.assertTrue(len(subtitles.keys()) == 0)
	197
	198	def test_multiple_langs(self):
	199	self.DL.params['writesubtitles'] = True
	200	langs = ['es', 'fr', 'de']
	201	self.DL.params['subtitleslangs'] = langs
	202	subtitles = self.getSubtitles()
	203	for lang in langs:
	204	self.assertTrue(subtitles.get(lang) is not None, u'Subtitles for \'%s\' not extracted' % lang)
	205
b4bcffef PH	206
	207	class TestBlipTVSubtitles(BaseTestSubtitles):
	208	url = 'http://blip.tv/a/a-6603250'
	209	IE = BlipTVIE
	210
	211	def test_list_subtitles(self):
	212	self.DL.expect_warning(u'Automatic Captions not supported by this server')
	213	self.DL.params['listsubtitles'] = True
	214	info_dict = self.getInfoDict()
	215	self.assertEqual(info_dict, None)
	216
	217	def test_allsubtitles(self):
	218	self.DL.expect_warning(u'Automatic Captions not supported by this server')
	219	self.DL.params['writesubtitles'] = True
	220	self.DL.params['allsubtitles'] = True
	221	subtitles = self.getSubtitles()
	222	self.assertEqual(set(subtitles.keys()), set(['en']))
	223	self.assertEqual(md5(subtitles['en']), '5b75c300af65fe4476dff79478bb93e4')
	224
	225
ae84f879 JMF	226	if __name__ == '__main__':
ae84f879 JMF	227	unittest.main()