learningai's picture
Update utils.py
66b0cbb
import os
from pathlib import Path
from logger import logging
from typing import Optional
import whisper
import config
from datetime import datetime
def get_video_from_yt(video_url : str, save_file_dir : str, video_name : str="yt_audio") -> Optional[Path]:
"""
Download YouTube video as an audio file in .wav . Returns the path of the
downloaded file as string
"""
logging.info(f"Attempting youtube video download for URL : {video_url}")
try :
# create directory if not exists
os.makedirs(save_file_dir, exist_ok=True)
filepath = f"{save_file_dir}/{video_name}_{datetime.now().strftime('%d%m%Y__%H%M%S')}"
# download the file
os.system(f'yt-dlp --quiet -o {filepath} -x --audio-format "wav" {video_url}')
if os.path.exists(f"{filepath}.wav"):
logging.info(f"Download successful. Audio file path : {filepath}.wav")
return f"{filepath}.wav"
else:
logging.info("Download unsuccessful. Please check logs.")
return None
except Exception as e:
logging.info("Download unsuccessful.")
logging.exception(e)
return None
def get_text_from_audio(audio_path : str) -> Optional[str]:
"""
Extracts text from audio file.
"""
logging.info(f"Attempting to extract text from : {audio_path}")
try :
model = whisper.load_model(name='tiny', download_root=config.MODEL_DIR)
results = model.transcribe(audio_path)
logging.info("Extraction successful.")
return results['text']
except Exception as e:
logging.info("Extraction failed.")
logging.exception(e)
return None
#URL = "https://www.youtube.com/watch?v=iO5LjrQaN9s"
#save_file_dir = "audio_files"
#video_path = get_video_from_yt(URL, save_file_dir)
#if video_path:
# print(get_text_from_audio(video_path))