audio_processor.py

import shutil
import os
import requests
import uuid
from urllib.parse import urlparse
from fastapi import UploadFile, HTTPException

UPLOAD_DIR = "uploads"
os.makedirs(UPLOAD_DIR, exist_ok=True)

async def download_audio_from_url(url: str) -> str:
    try:
        response = requests.get(url, stream=True)
        response.raise_for_status()
        
        # Parse URL to get path and extension, ignoring query params
        parsed_url = urlparse(url)
        path = parsed_url.path
        ext = os.path.splitext(path)[1]
        
        if not ext:
            ext = ".mp3" # Default to mp3 if no extension found
            
        # Generate a safe filename
        filename = f"{uuid.uuid4()}{ext}"
        
        file_location = f"{UPLOAD_DIR}/{filename}"
        with open(file_location, "wb") as f:
            for chunk in response.iter_content(chunk_size=8192):
                 f.write(chunk)
        return file_location
    except Exception as e:
        raise HTTPException(status_code=400, detail=f"Failed to download audio: {str(e)}")

async def process_audio(file: UploadFile = None, audio_url: str = None):
    if file:
        file_location = f"{UPLOAD_DIR}/{file.filename}"
        with open(file_location, "wb+") as file_object:
            shutil.copyfileobj(file.file, file_object)
        filename = file.filename
    elif audio_url:
        file_location = await download_audio_from_url(audio_url)
        filename = os.path.basename(file_location)
    else:
        raise HTTPException(status_code=400, detail="No audio file or URL provided")
    
    # In a real scenario, use librosa or pydub to get duration/sample_rate
    # For now, just return file size as a proxy for 'processed' metadata
    file_size = os.path.getsize(file_location)
    
    return {
        "filename": filename,
        "size_bytes": file_size,
        "format": filename.split(".")[-1],
        "duration_seconds": 0.0 # Placeholder
    }