def write_into_subtitle(response, output_path):
print('Waiting for writing subtitle to complete...')
# read the chinese punctuation
with codecs.open(output_path + 'transcript-text.txt', 'r',
'utf-8') as reader:
words = reader.read()
punc_index_list = []
punc_index = 0
for w in words:
if not w.isalpha() and w not in string.whitespace:
punc_index_list.append(punc_index)
punc_index += 1
elif w.isalpha():
punc_index += 1
with codecs.open(output_path + 'subtitle-no-punctuation.srt', 'w',
'utf-8') as writer:
i = 1 # setting the sequence number for srt
init = True # init flag
word_index = 0
curr = 0 # current punctuation number
for result in response.results:
alternative = result.alternatives[0]
line = "" # each line contain 10 words
counter = 0 # word counter in a line
# how many words remaining in this result
num_woeds = len(alternative.words)
start_next_para = True
# loop the word in the result
for word_info in alternative.words:
word_index += 1
num_woeds -= 1
counter += 1
word = word_info.word
if init:
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
init = False
if start_next_para:
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
start_next_para = False
if counter < 10:
# when the num of word in this line less than
# 10 word, we only add this word in this line
line += word
if word_index == (punc_index_list[curr]):
curr += 1
line += ' '
word_index += 1
else:
# the line is enouge 10 words, we inster seq num,
# time and line into the srt file
counter = 0 # clear the counter for nex iteration
end_time = word_info.end_time
str_end = timestr.timefm(end_time.seconds +
end_time.nanos * 1e-9)
writer.write(str(i)) # write the seq num into file,
# and then add 1
i += 1
line += word
if word_index == (punc_index_list[curr]):
curr += 1
line += ' '
word_index += 1
writer.write('\n')
writer.write(str_start) # write start time
writer.write(' --> ')
writer.write(str_end) # write end time
writer.write('\n')
writer.write(line) # write the word
line = "" # clear the line for next iteration
writer.write('\n\n')
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
# avoid miss any word, because counter < 0,
# but this iteration has no word remain
if counter < 10 and num_woeds == 0:
end_time = word_info.end_time
str_end = timestr.timefm(end_time.seconds +
end_time.nanos * 1e-9)
writer.write(str(i))
i += 1
writer.write('\n')
writer.write(str_start) # write start time
writer.write(' --> ')
writer.write(str_end) # write end time
writer.write('\n')
writer.write(line) # write the word
line = ""
writer.write('\n\n')
def write_into_subtitle(response, output_path):
print('Waiting for writing subtitle to complete...')
with codecs.open(output_path + 'subtitle-raw.srt', 'w', 'utf-8') as writer:
i = 1 # setting the sequence number for srt
init = True # init flag
for result in response.results:
alternative = result.alternatives[0]
line = "" # each line contain 10 words
counter = 0 # word counter in a line
# how many words remaining in this result
num_woeds = len(alternative.words)
start_next_para = True
# loop the word in the result
for word_info in alternative.words:
num_woeds -= 1
counter += 1
word = word_info.word
if init:
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
init = False
if start_next_para:
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
start_next_para = False
if counter < 10:
# when the num of word in this line less than
# 10 word, we only add this word in this line
line += word
else:
# the line is enouge 10 words, we inster seq num,
# time and line into the srt file
counter = 0 # clear the counter for nex iteration
end_time = word_info.end_time
str_end = timestr.timefm(end_time.seconds +
end_time.nanos * 1e-9)
writer.write(str(i)) # write the seq num into file,
# and then add 1
i += 1
line += word
writer.write('\n')
writer.write(str_start) # write start time
writer.write(' --> ')
writer.write(str_end) # write end time
writer.write('\n')
writer.write(line) # write the word
line = "" # clear the line for next iteration
writer.write('\n\n')
start_time = word_info.start_time
str_start = timestr.timefm(start_time.seconds +
start_time.nanos * 1e-9)
# avoid miss any word, because counter < 0,
# but this iteration has no word remain
if counter < 10 and num_woeds == 0:
end_time = word_info.end_time
str_end = timestr.timefm(end_time.seconds +
end_time.nanos * 1e-9)
writer.write(str(i))
i += 1
writer.write('\n')
writer.write(str_start) # write start time
writer.write(' --> ')
writer.write(str_end) # write end time
writer.write('\n')
writer.write(line) # write the word
line = ""
writer.write('\n\n')