Zulko / moviepy

Video editing with Python
https://zulko.github.io/moviepy/
MIT License
12.64k stars 1.59k forks source link

Writing Audio Gest Stuck! #2091

Open fdefake opened 10 months ago

fdefake commented 10 months ago

Hello there! I am using a script to generate subtitles for videos, it extracts the audio from the video and then subtitles it. My problem is the audio extraction, check this out: This is the code that generates subtitles, it needs to be called and passed the route of the video as an argument:

import os
import argparse
from datetime import timedelta
from moviepy.editor import VideoFileClip
import whisper
from moviepy.config import change_settings

change_settings({"FFMPEG_THREADS": 5})

def extract_audio(video, temp_file):
    if video.audio is not None:
        video.audio.write_audiofile(temp_file, codec="mp3")
    else:
        raise ValueError("Video has no audio.")

def generate_subtitles(video_path, srt_name, transcribe_segments, temp_file):
    video_dir, video_filename = os.path.split(video_path)
    file_name, _ = os.path.splitext(video_filename)

    srt_path = os.path.join(video_dir, srt_name)

    print('[i] Generating SRT...')
    with open(srt_path, "w", encoding="utf-8") as f:
        for segment_id, seg in enumerate(transcribe_segments, start=1):
            start = format_time(seg["start"])
            end = format_time(seg["end"])
            text = seg["text"].lstrip()
            f.write(f"{segment_id}\n{start} --> {end}\n{text}\n\n")
    try:
        os.remove(temp_file)
    except FileNotFoundError:
        print(f"Warning: Temporary MP3 file '{temp_file}' could not be found for deletion.")

def format_time(seconds):
    return str(timedelta(seconds=int(seconds))) + ",000"

def main():
    parser = argparse.ArgumentParser(description="auto caption generator v1.0")
    parser.add_argument("-f", "--video-path", type=str, required=True, help="Filepath of the video")
    parser.add_argument("--model", choices=["tiny", "base", "small", "medium", "large"], default="tiny", help="Choose the whisper model size (default: tiny)")
    args = parser.parse_args()
    path = args.video_path

    try:
        validate_path(path)
        video_manager = VideoFileClip(path)
        file_name, _ = os.path.splitext(os.path.basename(path))
        temp_file = f"{file_name}_temp_for_srt.mp3"
        srt_name = f"{file_name}.srt"

        with video_manager:
            extract_audio(video_manager, temp_file)
            model = whisper.load_model(args.model)
            transcribe = model.transcribe(audio=temp_file, fp16=False)
            print('[i] Transcription finished.')
            generate_subtitles(path, srt_name, transcribe["segments"], temp_file)
    except FileNotFoundError:
        print(f"Error: Temporary file '{temp_file}' not found.")
    except Exception as e:
        print(f"Error: {e}")

def validate_path(path):
    if not os.path.exists(path):
        raise ValueError("Invalid file path, quitting")

if __name__ == "__main__":
    main()

This code calls the script + the route of every video on a folder:

import os
import subprocess
from natsort import natsorted  # Asegúrate de tener instalado este módulo: pip install natsort

def process_videos(video_folder):
    videos = natsorted([file for file in os.listdir(video_folder) if file.endswith(".mp4")])

    for video in videos:
        video_path = os.path.join(video_folder, video)
        command = f"python auto_caption_generator.py -f \"{video_path}\" --model small"

        print(f"Processing video: {video}")
        process = subprocess.Popen(command, shell=True)
        process.communicate()

        print(f"Finished processing {video}")

if __name__ == "__main__":
    video_folder_path = "D:\BTO\MostlyToons\CRUD-VIDS-SubsEn"  # Reemplaza con tu ruta de carpeta
    process_videos(video_folder_path)

And this is the output:

Processing video: Batman T.A.S - S01 E25 - The Cape and Cowl Conspiracy (1080p - BluRay).mp4
MoviePy - Writing audio in Batman T.A.S - S01 E25 - The Cape and Cowl Conspiracy (1080p - BluRay)_temp_for_srt.mp3
chunk:  19%|█████████████████▍                                                                            | 5507/29715 [00:30<00:48, 498.47it/s, now=None]

It just gets stuck! Help Please, I am a begginer I do not get most of the things that happen over here!

SohamTilekar commented 10 months ago

Dose It Remains Stuck or Not. Plese provide the module Version, python version, etc.