def split_and_save_files(t, target_dir, subtitle_file):
hash = get_hash(subtitle_file + t["original_phrase"] + str(t["ts_start"]))
wav_file_dir = os.path.join(target_dir, "wav", hash[:2])
txt_file_dir = os.path.join(target_dir, "txt", hash[:2])
metadata_dir = os.path.join(target_dir, "metadata", hash[:2])
os.makedirs(wav_file_dir, exist_ok=True)
os.makedirs(txt_file_dir, exist_ok=True)
os.makedirs(metadata_dir, exist_ok=True)
target_wav_file = os.path.join(wav_file_dir, hash + ".wav")
target_txt_file = os.path.join(txt_file_dir, hash + ".txt")
target_metadata_file = os.path.join(metadata_dir, hash + ".json")
text = t["original_phrase"]
if len(text) == 0:
return
if not os.path.exists(target_wav_file) or not os.path.exists(
target_txt_file):
extract_audio_part_segment(audio_file, t["ts_start"], t["ts_end"],
target_wav_file)
with io.open(target_txt_file, "w", encoding='utf-8') as f:
f.write(text)
with io.open(target_metadata_file, "w", encoding='utf-8') as f:
t["ts_start"] = str(t["ts_start"])
t["ts_end"] = str(t["ts_end"])
t["metadata"] = metadata
json.dump(t, f)
assert os.path.exists(target_txt_file) and os.path.exists(target_wav_file) \
and getsize(target_wav_file) > 4 * 1024, "{} not created".format(target_wav_file)
def _get_transcript_google_web_asr(t):
import tempfile
try:
with tempfile.NamedTemporaryFile(suffix=".wav") as f:
extract_audio_part_segment(t["video_file"], t["ts_start"], t["ts_end"], f.name)
r = sr.Recognizer()
with sr.AudioFile(f.name) as source:
audio = r.record(source)
return r.recognize_google(audio)
except Exception as e:
print(e)
return None
metadata_dir = os.path.join(target_dir, "metadata", hash[:2])
os.makedirs(wav_file_dir, exist_ok=True)
os.makedirs(txt_file_dir, exist_ok=True)
os.makedirs(metadata_dir, exist_ok=True)
target_wav_file = os.path.join(wav_file_dir, hash + ".wav")
target_txt_file = os.path.join(txt_file_dir, hash + ".txt")
target_metadata_file = os.path.join(metadata_dir, hash + ".json")
text = t["original_phrase"]
if len(text) == 0:
continue
if not os.path.exists(target_wav_file) or not os.path.exists(
target_txt_file):
extract_audio_part_segment(video_file, t["ts_start"],
t["ts_end"], target_wav_file)
with io.open(target_txt_file, "w") as f:
f.write(text)
with io.open(target_metadata_file, "w") as f:
t["ts_start"] = str(t["ts_start"])
t["ts_end"] = str(t["ts_end"])
t["metadata"] = metadata
json.dump(t, f)
assert os.path.exists(target_txt_file) and os.path.exists(target_wav_file) \
and getsize(target_wav_file) > 4 * 1024, "{} not created".format(target_wav_file)
except Exception as e:
print("error : " + str(e))
termcolor.cprint(e, color="red")