def test_sbv_get_caption_text_multiline(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[2].text,
'Caption text #3 (line 1)\nCaption text #3 (line 2)')
self.assertListEqual(
vtt.captions[2].lines,
['Caption text #3 (line 1)', 'Caption text #3 (line 2)'])
def parse_sbv(sbv_path):
"""
To get around the webvtt library not supporting empty subtitles
"""
data = None
with open(sbv_path, "r", encoding="utf-8") as f:
data = f.readlines()
data_out = []
lines = []
for d in data:
if d != "\n":
lines.append(d)
continue
if len(lines) >= 2:
data_out.extend(lines)
data_out.append(d)
lines = []
if len(lines) >= 2:
data_out.extend(lines)
out_path = path.join(TMP_DIR, path.basename(sbv_path) + ".tmp")
with open(out_path, "w", encoding="utf-8") as f:
for d in data_out:
f.write(d)
return webvtt.from_sbv(out_path)
def test_sbv_conversion(self):
os.makedirs(OUTPUT_DIR)
copy(self._get_file('two_captions.sbv'), OUTPUT_DIR)
vtt = webvtt.from_sbv(os.path.join(OUTPUT_DIR, 'two_captions.sbv'))
vtt.save()
self.assertTrue(
os.path.exists(os.path.join(OUTPUT_DIR, 'two_captions.vtt')))
with open(os.path.join(OUTPUT_DIR, 'two_captions.vtt'),
'r',
encoding='utf-8') as f:
lines = [line.rstrip() for line in f.readlines()]
expected_lines = [
'WEBVTT',
'',
'00:00:00.378 --> 00:00:11.378',
'Caption text #1',
'',
'00:00:11.378 --> 00:00:12.305',
'Caption text #2 (line 1)',
'Caption text #2 (line 2)',
]
self.assertListEqual(lines, expected_lines)
def get_subtitle_file(filename: str) -> WebVTT:
file_ext = os.path.splitext(filename)[1]
if file_ext == ".srt":
return webvtt.from_srt(filename)
elif file_ext == ".sbv":
return webvtt.from_sbv(filename)
elif file_ext == ".vtt":
return webvtt.read(filename)
else:
raise ValueError(filename)
def test_sbv_get_caption_text_multiline(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(
vtt.captions[2].text,
'Caption text #3 (line 1)\nCaption text #3 (line 2)'
)
self.assertListEqual(
vtt.captions[2].lines,
['Caption text #3 (line 1)', 'Caption text #3 (line 2)']
)
def sbv2df(sbv,textCol):
"""
Store (start, end, and text) of each time segment in the sbv file in a row of a pandas dataframe.
Input args
sbv (string): the file path of an sbv file
textCol (string): the name of the text column
"""
data = []
global webvtt
webvtt = webvtt.from_sbv(sbv)
for caption in webvtt:
data.append({'start':datetime.strptime(caption.start,'%H:%M:%S.%f').time(),
'end':datetime.strptime(caption.end,'%H:%M:%S.%f').time(),
textCol:caption.text})
df = pd.DataFrame(data)
df = df.replace('\n',' ', regex=True)
df = df[['start','end',textCol]]
return df
def fix_subtitle_sequencing(filename):
if os.path.isfile(filename + ".bk"):
print("Not overwriting original backup for {}, skipping.".format(
filename))
return
subs = None
if os.path.splitext(filename)[1] == ".srt":
subs = webvtt.from_srt(filename)
elif os.path.splitext(filename)[1] == ".sbv":
subs = webvtt.from_sbv(filename)
# Adjust timing and stretch subtitles for fixing the live ones which
# get messed up by Youtube
if "--fix-live" in sys.argv:
for i in range(len(subs)):
start = parse_time_stamp(subs[i].start)
start -= timedelta(seconds=8)
if start < timedelta(hours=0, minutes=0, seconds=0,
milliseconds=0):
start = timedelta(hours=0,
minutes=0,
seconds=0,
milliseconds=0)
end = start + timedelta(seconds=4)
subs[i].start = format_time_stamp(start)
subs[i].end = format_time_stamp(end)
for i in range(len(subs) - 1):
end = parse_time_stamp(subs[i].end)
next_start = parse_time_stamp(subs[i + 1].start)
if end > next_start:
subs[i].end = subs[i + 1].start
if not "--dry" in sys.argv:
shutil.copy(filename, filename + ".bk")
out_srt = os.path.splitext(filename)[0] + ".srt"
with open(out_srt, "w", encoding="utf8") as f:
subs.write(f, format="srt")
def test_sbv_conversion(self):
os.makedirs(OUTPUT_DIR)
copy(self._get_file('two_captions.sbv'), OUTPUT_DIR)
vtt = webvtt.from_sbv(os.path.join(OUTPUT_DIR, 'two_captions.sbv'))
vtt.save()
self.assertTrue(os.path.exists(os.path.join(OUTPUT_DIR, 'two_captions.vtt')))
with open(os.path.join(OUTPUT_DIR, 'two_captions.vtt'), 'r', encoding='utf-8') as f:
lines = [line.rstrip() for line in f.readlines()]
expected_lines = [
'WEBVTT',
'',
'00:00:00.378 --> 00:00:11.378',
'Caption text #1',
'',
'00:00:11.378 --> 00:00:12.305',
'Caption text #2 (line 1)',
'Caption text #2 (line 2)',
]
self.assertListEqual(lines, expected_lines)
def test_sbv_timestamps_in_seconds(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].start_in_seconds, 11.378)
self.assertEqual(vtt.captions[1].end_in_seconds, 12.305)
def test_sbv_missing_caption_text(self):
self.assertTrue(
webvtt.from_sbv(
self._get_file('missing_caption_text.sbv')).captions)
def test_sbv_get_caption_text(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].text, 'Caption text #2')
def test_sbv_timestamps_in_seconds(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].start_in_seconds, 11.378)
self.assertEqual(vtt.captions[1].end_in_seconds, 12.305)
def test_sbv_timestamps_format(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].start, '00:00:11.378')
self.assertEqual(vtt.captions[1].end, '00:00:12.305')
def test_sbv_total_length(self):
self.assertEqual(
webvtt.from_sbv(self._get_file('sample.sbv')).total_length,
16
)
def test_sbv_get_caption_text(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].text, 'Caption text #2')
def test_sbv_timestamps_format(self):
vtt = webvtt.from_sbv(self._get_file('sample.sbv'))
self.assertEqual(vtt.captions[1].start, '00:00:11.378')
self.assertEqual(vtt.captions[1].end, '00:00:12.305')
def test_sbv_total_length(self):
self.assertEqual(
webvtt.from_sbv(self._get_file('sample.sbv')).total_length, 16)