app.py

import os
import subprocess
import sys

# Ensure required libraries are installed
try:
    import yt_dlp as youtube_dl
except ImportError:
    subprocess.check_call([sys.executable, "-m", "pip", "install", "yt-dlp"])
    import yt_dlp as youtube_dl

import gradio as gr
from transformers import pipeline

def download_video(video_url, filename="downloaded_video.mp4"):
    """
    Download the video file using yt_dlp with cookies.txt for authentication.
    """
    ydl_opts = {
        'format': 'best',
        'outtmpl': filename,
        'noplaylist': True,
        'quiet': True,
        'cookiefile': 'cookies.txt',
        'user_agent': (
            'Mozilla/5.0 (Windows NT 10.0; Win64; x64) '
            'AppleWebKit/537.36 (KHTML, like Gecko) '
            'Chrome/115.0.0.0 Safari/537.36'
        ),
    }

    with youtube_dl.YoutubeDL(ydl_opts) as ydl:
        ydl.download([video_url])
    return filename

def extract_audio(video_filename, audio_filename="extracted_audio.wav"):
    command = [
        'ffmpeg',
        '-y',
        '-i', video_filename,
        '-vn',
        '-acodec', 'pcm_s16le',
        '-ar', '44100',
        '-ac', '2',
        audio_filename
    ]
    subprocess.run(command, check=True)
    return audio_filename

def classify_accent(audio_file, model_name="your-hf-accent-model"):
    if model_name == "your-hf-accent-model":
        return "Accent: Demo Accent\nConfidence: 100.00%"
    classifier = pipeline("audio-classification", model=model_name)
    results = classifier(audio_file)
    if results:
        top = results[0]
        return f"Accent: {top['label']}\nConfidence: {top['score'] * 100:.2f}%"
    else:
        return "No classification result."

def accent_classifier(video_url):
    try:
        video_filename = download_video(video_url)
        audio_filename = extract_audio(video_filename)
        result = classify_accent(audio_filename)
    except Exception as e:
        result = f"Error occurred: {e}"
    finally:
        if os.path.exists("downloaded_video.mp4"):
            os.remove("downloaded_video.mp4")
        if os.path.exists("extracted_audio.wav"):
            os.remove("extracted_audio.wav")
    return result

iface = gr.Interface(
    fn=accent_classifier,
    inputs=gr.Textbox(label="Video URL", placeholder="Enter a public YouTube, Vimeo, or Dailymotion link"),
    outputs="text",
    title="Accent Classifier",
    description="Downloads a video, extracts audio, and classifies the accent using a Hugging Face model."
)

if __name__ == "__main__":
    iface.launch()