app.py

from fastapi import FastAPI, UploadFile, File, HTTPException
from fastapi.responses import PlainTextResponse
from fastapi.middleware.gzip import GZipMiddleware
from recognizer_tunisian_vosk import RecognizerTunisianVosk
from pydub import AudioSegment
import shutil
import os

app = FastAPI()
app.add_middleware(GZipMiddleware, minimum_size=1000)

vosk_recognizer = RecognizerTunisianVosk()

TEMP_RAW = "temp_input"
TEMP_WAV = "temp.wav"

@app.get("/")
def read_root():
    return {"message": "Audio Transcription API is running."}

@app.post("/transcribe/tunisian", response_class=PlainTextResponse)
async def transcribe_tunisian(file: UploadFile = File(...)):
    try:
        # Save uploaded file
        with open(TEMP_RAW, "wb") as buffer:
            shutil.copyfileobj(file.file, buffer)

        # Convert to correct format using pydub
        audio = AudioSegment.from_file(TEMP_RAW)
        audio = audio.set_channels(1).set_frame_rate(16000)
        audio.export(TEMP_WAV, format="wav")

        # Transcribe
        text = vosk_recognizer.transcribe(TEMP_WAV)
        return text

    except Exception as e:
        raise HTTPException(status_code=400, detail=f"Audio processing error: {str(e)}")

    finally:
        for path in [TEMP_RAW, TEMP_WAV]:
            if os.path.exists(path):
                os.remove(path)