app.py

import gradio as gr
from moviepy.editor import VideoFileClip
from speechbrain.pretrained import EncoderClassifier
import torchaudio
from pytube import YouTube
import os

CLASSIFIER = "Jzuluaga/accent-id-commonaccent_xlsr-en-english"

def download_video(url):
    """Handles YouTube and direct video links"""
    if "youtube.com" in url:
        yt = YouTube(url)
        stream = yt.streams.filter(progressive=True, file_extension='mp4').first()
        video_path = stream.download()
        return video_path
    else:  # Direct download (assumes URL is direct mp4 link)
        # Download file locally
        import requests
        local_filename = "temp_video.mp4"
        with requests.get(url, stream=True) as r:
            r.raise_for_status()
            with open(local_filename, 'wb') as f:
                for chunk in r.iter_content(chunk_size=8192):
                    f.write(chunk)
        return local_filename

def extract_audio(video_path):
    clip = VideoFileClip(video_path)
    audio_path = "temp_audio.wav"
    clip.audio.write_audiofile(audio_path, logger=None)
    clip.close()
    return audio_path

def classify_accent(audio_path):
    classifier = EncoderClassifier.from_hparams(
        source=CLASSIFIER,
        savedir="pretrained_models/accent_classifier",
        run_opts={"device":"cpu"}  # or "cuda" if GPU available
    )
    waveform, sample_rate = torchaudio.load(audio_path)
    prediction = classifier.classify_batch(waveform)
    # prediction format: (scores, probabilities, embeddings, predicted_labels)
    predicted_accent = prediction[3][0]
    confidence = prediction[1].exp().max().item() * 100
    return predicted_accent, f"{confidence:.2f}%"

def process_video(url):
    try:
        video_path = download_video(url)
        audio_path = extract_audio(video_path)
        accent, confidence = classify_accent(audio_path)
    finally:
        # Cleanup temp files if they exist
        for f in [video_path, audio_path]:
            if os.path.exists(f):
                os.remove(f)
    return accent, confidence

# Gradio interface
iface = gr.Interface(
    fn=process_video,
    inputs=gr.Textbox(label="Enter Public Video URL (YouTube, Loom, direct MP4)"),
    outputs=[
        gr.Textbox(label="Detected Accent"),
        gr.Textbox(label="Confidence Score")
    ],
    title="English Accent Classifier",
    description="Paste a public video URL to detect the English accent and confidence score."
)

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