experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

simplify function

Browse Source
This commit is contained in:
JRMeyer 2020-04-02 21:40:57 -07:00
parent aa223d7090
commit cad03d33ea
1 changed files with 8 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
transcribe.py
View File

@ -76,20 +76,13 @@ def transcribe_file(audio_path, tlog_path):
json.dump(transcripts, tlog_file, default=float) json.dump(transcripts, tlog_file, default=float)
def transcribe_many(paths, kind): def transcribe_many(src_paths,dst_paths):
pbar = create_progressbar(prefix='Transcribing files | ', max_value=len(paths)).start() pbar = create_progressbar(prefix='Transcribing files | ', max_value=len(src_paths)).start()
if kind == 'dir': for i in range(len(src_paths)):
# the user pointed to a dir of files
src_paths = paths
dst_paths = [ path.replace('.wav','.tlog') for path in paths ]
elif kind == 'catalog':
# the user pointed to a catalog dir from DSAlign
src_paths,dst_paths = zip(*paths)
for i in range(len(paths)):
p = Process(target=transcribe_file, args=(src_paths[i], dst_paths[i])) p = Process(target=transcribe_file, args=(src_paths[i], dst_paths[i]))
p.start() p.start()
p.join() p.join()
log_progress('Transcribed file {} of {} from "{}" to "{}"'.format(i + 1, len(paths), src_paths[i], dst_paths[i])) log_progress('Transcribed file {} of {} from "{}" to "{}"'.format(i + 1, len(src_paths), src_paths[i], dst_paths[i]))
pbar.update(i) pbar.update(i)
pbar.finish() pbar.finish()
@ -129,7 +122,8 @@ def main(_):
fail('Destination file(s) from catalog already existing, use --force for overwriting') fail('Destination file(s) from catalog already existing, use --force for overwriting')
if any(map(lambda e: not os.path.isdir(os.path.dirname(e[1])), catalog_entries)): if any(map(lambda e: not os.path.isdir(os.path.dirname(e[1])), catalog_entries)):
fail('Missing destination directory for at least one catalog entry') fail('Missing destination directory for at least one catalog entry')
transcribe_many(catalog_entries, 'catalog') src_paths,dst_paths = zip(*paths)
transcribe_many(src_paths,dst_paths)
else: else:
# Transcribe one file # Transcribe one file
dst_path = os.path.abspath(FLAGS.dst) if FLAGS.dst else os.path.splitext(src_path)[0] + '.tlog' dst_path = os.path.abspath(FLAGS.dst) if FLAGS.dst else os.path.splitext(src_path)[0] + '.tlog'
@ -153,7 +147,8 @@ def main(_):
wav_paths = glob.glob(src_path + "/*.wav") wav_paths = glob.glob(src_path + "/*.wav")
else: else:
wav_paths = glob.glob(src_path + "/**/*.wav") wav_paths = glob.glob(src_path + "/**/*.wav")
transcribe_many(wav_paths, 'dir') dst_paths = [path.replace('.wav','.tlog') for path in wav_paths]
transcribe_many(wav_paths,dst_paths)
if __name__ == '__main__': if __name__ == '__main__':