app.py

from fastapi import FastAPI, File, UploadFile, HTTPException
from fastapi.responses import JSONResponse
import os
import tempfile
import shutil
from typing import Optional
import threading
import uvicorn

os.makedirs("/tmp/transformers_cache", exist_ok=True)
os.makedirs("/tmp/hf_home", exist_ok=True)
os.makedirs("/tmp/torch_home", exist_ok=True)

os.environ["TRANSFORMERS_CACHE"] = "/tmp/transformers_cache"
os.environ["HF_HOME"] = "/tmp/hf_home"
os.environ["TORCH_HOME"] = "/tmp/torch_home"

app = FastAPI(
    title="Speech Transcription API",
    description="API for transcribing speech using Whisper model",
    version="1.0.0"
)

model_loaded = False
transcriber = None

def load_model():
    """Load the model in a background thread"""
    global transcriber, model_loaded
    try:
        from transformers import pipeline
        transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-small")
        model_loaded = True
        print("Model loaded successfully!")
    except Exception as e:
        print(f"Error loading model: {e}")

threading.Thread(target=load_model, daemon=True).start()

@app.get("/")
def read_root():
    return {"message": "Welcome to the Speech Transcription API. Use /transcribe endpoint to transcribe audio."}

@app.get("/health")
def health_check():
    """Health check endpoint"""
    return {"status": "healthy", "model_loaded": model_loaded}

@app.post("/transcribe")
async def transcribe_audio(file: UploadFile = File(...), 
                           return_format: Optional[str] = "json"):
    global model_loaded, transcriber

    if not model_loaded:
        raise HTTPException(
            status_code=503, 
            detail="Model is still loading, please try again in a few minutes"
        )

    if not file.filename.lower().endswith(('.wav', '.mp3', '.m4a', '.ogg')):
        raise HTTPException(
            status_code=400, 
            detail="Unsupported file format. Please upload WAV, MP3, M4A, or OGG file."
        )
    
    with tempfile.NamedTemporaryFile(delete=False, dir="/tmp") as temp_file:
        shutil.copyfileobj(file.file, temp_file)
        temp_path = temp_file.name
    
    try:
        result = transcriber(temp_path)
        transcript = result["text"]
        
        os.unlink(temp_path)

        if return_format and return_format.lower() == "text":
            return transcript
        else: 
            return JSONResponse(content={"transcript": transcript})
    
    except Exception as e:
        if os.path.exists(temp_path):
            os.unlink(temp_path)
        raise HTTPException(status_code=500, detail=f"Error processing audio: {str(e)}")

if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=7860)