def get_subs(vtt_subs_path):
subs = []
reader = WebVTTReader()
with open(vtt_subs_path, 'r') as f:
text = f.read().decode(utils.get_file_encoding(vtt_subs_path))
vtt = reader.read(text)
vttsubs = vtt.get_captions(vtt.get_languages()[0])
#vttsubs = pyvtt.WebVTTFile.open(vtt_subs_path)
print "vttsubs total: %i " % len(vttsubs)
print vttsubs[0].start
print vttsubs[0].end
print vttsubs[0].get_text()
for s in vttsubs:
subs.append({
"text": s.get_text(),
"start": float(s.start) / 1000000,
"end": float(s.end) / 1000000
})
return subs
def get_captions_from_output(self,
output: str,
language: str = 'en') -> str:
reader = WebVTTReader()
temp_final = []
for caption in reader.read(output, language).get_captions(language):
stripped = str(caption).split("\\n")[-1].replace("'", '')
timestamp = self.get_time_from_caption(str(caption))
temp_final.append(",".join([timestamp[0], timestamp[1], stripped]))
lst = [[], [], []]
for c in temp_final:
for item in range(len(c.split(","))):
lst[item].append(c.split(",")[item])
df = pd.DataFrame({'start': lst[0], 'end': lst[1], 'content': lst[2]})
# final = ''
# previous = ''
# for line in temp_final.split("\n"):
# if previous != line:
# final += "," + line
# previous = line
return df
class WebVTTReaderTestCase(unittest.TestCase):
def setUp(self):
self.reader = WebVTTReader()
def test_positive_answer_for_detection(self):
self.assertTrue(self.reader.detect(SAMPLE_WEBVTT))
def test_negative_answer_for_detection(self):
self.assertFalse(self.reader.detect(SAMPLE_SRT))
def test_caption_length(self):
captions = self.reader.read(SAMPLE_WEBVTT)
self.assertEqual(len(captions.get_captions('en-US')), 7)
def test_read_supports_multiple_languages(self):
captions = self.reader.read(SAMPLE_WEBVTT, lang='es')
self.assertIsNotNone(captions.get_captions('es'))
def test_proper_timestamps(self):
captions = self.reader.read(SAMPLE_WEBVTT)
cue = captions.get_captions('en-US')[2]
self.assertEqual(cue.start, 17000000)
self.assertEqual(cue.end, 18752000)
def test_webvtt_cue_components_removed_from_text(self):
result = self.reader._clean(
"\n" # the first line is sckipped by the cleaner
"<c vIntro><b>Wikipedia</b> is a great adventure. <i>It may have "
"its shortcomings</i>, but it is<u> the largest</u> collective "
"knowledge construction endevour</c> <ruby>base text <rt>"
"annotation</rt></ruby> <v Audry><b>Yes</b>, indeed!"
)
expected = (
"Wikipedia is a great adventure. It may have "
"its shortcomings, but it is the largest collective "
"knowledge construction endevour base text annotation"
" Audry: Yes, indeed!"
)
self.assertEqual(result, expected)
def test_empty_file(self):
self.assertRaises(
CaptionReadNoCaptions,
WebVTTReader().read, SAMPLE_WEBVTT_EMPTY)
def get_captions_from_output(self, output: str, url: str) -> str:
reader = WebVTTReader()
captions = []
for caption in reader.read(output).get_captions('en-US'):
stripped = self.remove_time_from_caption(
url,
str(caption).replace(r'\n', " "))
stripped += "\n"
captions.append(stripped)
if self.search_query == '':
return ''.join(item for item in captions)
return self.process_captions(captions, url)
def get_captions_from_output(self, output: str) -> str:
reader = WebVTTReader()
temp_final = ''
for caption in reader.read(output).get_captions('en-US'):
stripped = self.remove_time_from_caption(
str(caption).replace(r'\n', "\n"))
temp_final += stripped
final = ''
previous = ''
for line in temp_final.split("\n"):
if previous != line:
final += "\n" + line
previous = line
return final.replace("\n", ' ')[1:]
class WebVTTReaderTestCase(unittest.TestCase):
def setUp(self):
self.reader = WebVTTReader()
def test_positive_answer_for_detection(self):
self.assertTrue(self.reader.detect(SAMPLE_WEBVTT.decode(u'utf-8')))
def test_negative_answer_for_detection(self):
self.assertFalse(self.reader.detect(SAMPLE_SRT.decode(u'utf-8')))
def test_caption_length(self):
captions = self.reader.read(SAMPLE_WEBVTT_2.decode(u'utf-8'))
self.assertEqual(len(captions.get_captions(u'en-US')), 7)
def test_read_supports_multiple_languages(self):
captions = self.reader.read(SAMPLE_WEBVTT.decode(u'utf-8'), lang=u'es')
self.assertIsNotNone(captions.get_captions(u'es'))
def test_proper_timestamps(self):
captions = self.reader.read(SAMPLE_WEBVTT.decode(u'utf-8'))
cue = captions.get_captions(u'en-US')[2]
self.assertEqual(cue.start, 17000000)
self.assertEqual(cue.end, 18752000)
def test_webvtt_cue_components_removed_from_text(self):
result = self.reader._remove_styles(
u"<c vIntro><b>Wikipedia</b> is a great adventure. <i>It may have "
u"its shortcomings</i>, but it is<u> the largest</u> collective "
u"knowledge construction endevour</c> <ruby>base text <rt>"
u"annotation</rt></ruby> <v Audry><b>Yes</b>, indeed!"
)
expected = (
u"Wikipedia is a great adventure. It may have "
u"its shortcomings, but it is the largest collective "
u"knowledge construction endevour base text annotation"
u" Audry: Yes, indeed!"
)
self.assertEqual(result, expected)
def test_empty_file(self):
self.assertRaises(
CaptionReadNoCaptions,
WebVTTReader().read, SAMPLE_WEBVTT_EMPTY.decode(u'utf-8'))
def test_not_ignoring_timing_errors(self):
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader(ignore_timing_errors=False).read,
(u"\n"
u"00:00:20,000 --> 00:00:10,000\n"
u"foo bar baz")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20,000 --> 00:00:10,000\n"
u"Start time is greater than end time.\n")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20,000 --> 00:00:30,000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10,000 --> 00:00:20,000\n"
u"This cue starts before the previous one.\n")
)
def test_ignoring_timing_errors(self):
# Even if timing errors are ignored, this is worse
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
(u"\nNOTE invalid cue stamp\n"
u"00:00:20,000 --> \n"
u"foo bar baz\n")
)
try:
WebVTTReader().read(
(u"\n"
u"00:00:20,000 --> 00:00:10,000\n"
u"Start time is greater than end time.\n")
)
except CaptionReadError:
self.fail(u"Shouldn't raise CaptionReadError")
try:
WebVTTReader().read(
(u"\n"
u"00:00:20,000 --> 00:00:30,000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10,000 --> 00:00:20,000\n"
u"This cue starts before the previous one.\n")
)
except CaptionReadError:
self.fail(u"Shouldn't raise CaptionReadError")
def test_invalid_files(self):
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
(u"\nNOTE Cues without text are invalid.\n"
u"00:00:20,000 --> 00:00:30,000\n"
u"\n"
u"00:00:40,000 --> 00:00:50,000\n"
u"foo bar baz\n")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20,000 --> 00:00:10,000\n"
u"Start time is greater than end time.")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20,000 --> 00:00:30,000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10,000 --> 00:00:20,000\n"
u"This cue starts before the previous one.\n")
)
class WebVTTReaderTestCase(unittest.TestCase):
def setUp(self):
self.reader = WebVTTReader()
def test_positive_answer_for_detection(self):
self.assertTrue(self.reader.detect(SAMPLE_WEBVTT.decode(u'utf-8')))
def test_negative_answer_for_detection(self):
self.assertFalse(self.reader.detect(SAMPLE_SRT.decode(u'utf-8')))
def test_caption_length(self):
captions = self.reader.read(SAMPLE_WEBVTT_2.decode(u'utf-8'))
self.assertEqual(len(captions.get_captions(u'en-US')), 7)
def test_read_supports_multiple_languages(self):
captions = self.reader.read(SAMPLE_WEBVTT.decode(u'utf-8'), lang=u'es')
self.assertIsNotNone(captions.get_captions(u'es'))
def test_proper_timestamps(self):
captions = self.reader.read(SAMPLE_WEBVTT.decode(u'utf-8'))
cue = captions.get_captions(u'en-US')[2]
self.assertEqual(cue.start, 17000000)
self.assertEqual(cue.end, 18752000)
def test_webvtt_cue_components_removed_from_text(self):
result = self.reader._remove_styles(
u"<c vIntro><b>Wikipedia</b> is a great adventure. <i>It may have "
u"its shortcomings</i>, but it is<u> the largest</u> collective "
u"knowledge construction endevour</c> <ruby>base text <rt>"
u"annotation</rt></ruby> <v Audry><b>Yes</b>, indeed!"
)
expected = (
u"Wikipedia is a great adventure. It may have "
u"its shortcomings, but it is the largest collective "
u"knowledge construction endevour base text annotation"
u" Audry: Yes, indeed!"
)
self.assertEqual(result, expected)
def test_empty_file(self):
self.assertRaises(
CaptionReadNoCaptions,
WebVTTReader().read, SAMPLE_WEBVTT_EMPTY.decode(u'utf-8'))
def test_invalid_files(self):
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
u"""
NOTE Cues without text are invalid.
00:00:20,000 --> 00:00:10,000
"""
)
self.assertRaises(
CaptionReadError,
WebVTTReader().read,
u"""
00:00:20,000 --> 00:00:10,000
Start time is greater than end time.
"""
)
self.assertRaises(
CaptionReadError,
WebVTTReader().read,
u"""
00:00:20,000 --> 00:00:30,000
Start times should be consecutive.
00:00:10,000 --> 00:00:20,000
This cue starts before the previous one.
"""
)
class WebVTTReaderTestCase(unittest.TestCase):
def setUp(self):
self.reader = WebVTTReader()
def test_positive_answer_for_detection(self):
self.assertTrue(self.reader.detect(SAMPLE_WEBVTT))
def test_negative_answer_for_detection(self):
self.assertFalse(self.reader.detect(SAMPLE_SRT))
def test_caption_length(self):
captions = self.reader.read(SAMPLE_WEBVTT_2)
self.assertEqual(len(captions.get_captions(u'en-US')), 7)
def test_read_supports_multiple_languages(self):
captions = self.reader.read(SAMPLE_WEBVTT, lang=u'es')
self.assertIsNotNone(captions.get_captions(u'es'))
def test_proper_timestamps(self):
captions = self.reader.read(SAMPLE_WEBVTT)
cue = captions.get_captions(u'en-US')[2]
self.assertEqual(cue.start, 17000000)
self.assertEqual(cue.end, 18752000)
def test_webvtt_cue_components_removed_from_text(self):
result = self.reader._remove_styles(
u"<c vIntro><b>Wikipedia</b> is a great adventure. <i>It may have "
u"its shortcomings</i>, but it is<u> the largest</u> collective "
u"knowledge construction endevour</c> <ruby>base text <rt>"
u"annotation</rt></ruby> <v Audry><b>Yes</b>, indeed!"
)
expected = (
u"Wikipedia is a great adventure. It may have "
u"its shortcomings, but it is the largest collective "
u"knowledge construction endevour base text annotation"
u" Audry: Yes, indeed!"
)
self.assertEqual(result, expected)
def test_empty_file(self):
self.assertRaises(
CaptionReadNoCaptions,
WebVTTReader().read, SAMPLE_WEBVTT_EMPTY)
def test_not_ignoring_timing_errors(self):
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"\n"
u"00:00:20.000 --> 00:00:10.000\n"
u"foo bar baz")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20.000 --> 00:00:10.000\n"
u"Start time is greater than end time.\n")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20.000 --> 00:00:30.000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10.000 --> 00:00:20.000\n"
u"This cue starts before the previous one.\n")
)
def test_ignoring_timing_errors(self):
# Even if timing errors are ignored, this has to raise an exception
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
(u"\nNOTE invalid cue stamp\n"
u"00:00:20.000 --> \n"
u"foo bar baz\n")
)
# And this too
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
(u"\n00:00:20,000 --> 00:00:22,000\n"
u"Note the comma instead of point.\n")
)
try:
WebVTTReader().read(
(u"\n"
u"00:00:20.000 --> 00:00:10.000\n"
u"Start time is greater than end time.\n")
)
except CaptionReadError:
self.fail(u"Shouldn't raise CaptionReadError")
try:
WebVTTReader().read(
(u"\n"
u"00:00:20.000 --> 00:00:30.000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10.000 --> 00:00:20.000\n"
u"This cue starts before the previous one.\n")
)
except CaptionReadError:
self.fail(u"Shouldn't raise CaptionReadError")
def test_invalid_files(self):
self.assertRaises(
CaptionReadSyntaxError,
WebVTTReader().read,
(u"\nNOTE Cues without text are invalid.\n"
u"00:00:20.000 --> 00:00:30.000\n"
u"\n"
u"00:00:40.000 --> 00:00:50.000\n"
u"foo bar baz\n")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20.000 --> 00:00:10.000\n"
u"Start time is greater than end time.")
)
self.assertRaises(
CaptionReadError,
WebVTTReader(ignore_timing_errors=False).read,
(u"00:00:20.000 --> 00:00:30.000\n"
u"Start times should be consecutive.\n"
u"\n"
u"00:00:10.000 --> 00:00:20.000\n"
u"This cue starts before the previous one.\n")
)
def test_zero_start(self):
captions = self.reader.read(SAMPLE_WEBVTT_LAST_CUE_ZERO_START)
cue = captions.get_captions(u'en-US')[0]
self.assertEquals(cue.start, 0)
class TestWebVTTReader:
def setup_method(self):
self.reader = WebVTTReader()
def test_positive_answer_for_detection(self, sample_webvtt):
assert self.reader.detect(sample_webvtt) is True
def test_negative_answer_for_detection(self, sample_srt):
assert self.reader.detect(sample_srt) is False
def test_caption_length(self, sample_webvtt_2):
captions = self.reader.read(sample_webvtt_2)
assert len(captions.get_captions('en-US')) == 7
def test_read_supports_multiple_languages(self, sample_webvtt):
captions = self.reader.read(sample_webvtt, lang='es')
assert captions.get_captions('es') is not None
def test_proper_timestamps(self, sample_webvtt):
captions = self.reader.read(sample_webvtt)
cue = captions.get_captions('en-US')[2]
assert cue.start == 17000000
assert cue.end == 18752000
def test_webvtt_cue_components_removed_from_text(self):
result = self.reader._remove_styles(
"<c vIntro><b>Wikipedia</b> is a great adventure. <i>It may have "
"its shortcomings</i>, but it is<u> the largest</u> collective "
"knowledge construction endevour</c> <ruby>base text <rt>"
"annotation</rt></ruby> <v Audry><b>Yes</b>, indeed!")
expected = ("Wikipedia is a great adventure. It may have "
"its shortcomings, but it is the largest collective "
"knowledge construction endevour base text annotation"
" Audry: Yes, indeed!")
assert result == expected
def test_empty_file(self, sample_webvtt_empty):
with pytest.raises(CaptionReadNoCaptions):
WebVTTReader().read(sample_webvtt_empty)
def test_not_ignoring_timing_errors(self):
# todo: same assert w/ different arguments -> this can be parametrized;
with pytest.raises(CaptionReadError):
WebVTTReader(ignore_timing_errors=False).read(
"\n"
"00:00:20.000 --> 00:00:10.000\n"
"foo bar baz")
with pytest.raises(CaptionReadError):
WebVTTReader(ignore_timing_errors=False).read(
"00:00:20.000 --> 00:00:10.000\n"
"Start time is greater than end time.\n")
with pytest.raises(CaptionReadError):
WebVTTReader(ignore_timing_errors=False).read(
"00:00:20.000 --> 00:00:30.000\n"
"Start times should be consecutive.\n"
"\n"
"00:00:10.000 --> 00:00:20.000\n"
"This cue starts before the previous one.\n")
def test_ignoring_timing_errors(self):
# Even if timing errors are ignored, this has to raise an exception
with pytest.raises(CaptionReadSyntaxError):
WebVTTReader().read(
"\nNOTE invalid cue stamp\n00:00:20.000 --> \nfoo bar baz\n")
# And this too
with pytest.raises(CaptionReadSyntaxError):
WebVTTReader().read("\n00:00:20,000 --> 00:00:22,000\n"
"Note the comma instead of point.\n")
# todo: at this point it can be split into 2 separate tests
try:
WebVTTReader().read("\n"
"00:00:20.000 --> 00:00:10.000\n"
"Start time is greater than end time.\n")
except CaptionReadError:
pytest.fail("Shouldn't raise CaptionReadError")
try:
WebVTTReader().read("\n"
"00:00:20.000 --> 00:00:30.000\n"
"Start times should be consecutive.\n"
"\n"
"00:00:10.000 --> 00:00:20.000\n"
"This cue starts before the previous one.\n")
except CaptionReadError:
pytest.fail("Shouldn't raise CaptionReadError")
def test_invalid_files(self):
with pytest.raises(CaptionReadError):
WebVTTReader(ignore_timing_errors=False).read(
"00:00:20.000 --> 00:00:10.000\n"
"Start time is greater than end time.")
with pytest.raises(CaptionReadError):
WebVTTReader(ignore_timing_errors=False).read(
"00:00:20.000 --> 00:00:30.000\n"
"Start times should be consecutive.\n"
"\n"
"00:00:10.000 --> 00:00:20.000\n"
"This cue starts before the previous one.\n")
def test_zero_start(self, sample_webvtt_last_cue_zero_start):
captions = self.reader.read(sample_webvtt_last_cue_zero_start)
cue = captions.get_captions('en-US')[0]
assert cue.start == 0
def test_webvtt_empty_cue(self, sample_webvtt_empty_cue):
assert 1 == len(
self.reader.read(sample_webvtt_empty_cue).get_captions('en-US'))
