app.py

from fastapi import FastAPI, UploadFile, Form
from fastapi.responses import FileResponse, JSONResponse
from piper.voice import PiperVoice
import soundfile as sf
import numpy as np
import uuid
import os

app = FastAPI()

# Load the voice model
voice_model_path = "en_US-hfc_female-medium.onnx"
voice = PiperVoice.load(voice_model_path)

# Directory to store generated audio files
output_dir = "generated_audio"
os.makedirs(output_dir, exist_ok=True)


@app.post("/text-to-speech/")
async def text_to_speech(text: str = Form(...)):
    """
    Convert text to speech and return the generated audio file.

    Args:
        text: The input text to convert to speech.

    Returns:
        The path to the generated audio file.
    """
    try:
        # Generate a unique filename
        output_file = os.path.join(output_dir, f"{uuid.uuid4()}.ogg")

        # Synthesize and save the audio
        with sf.SoundFile(output_file, mode='w', samplerate=voice.config.sample_rate,
                          channels=1, format='OGG', subtype='VORBIS') as ogg_file:
            for audio_bytes in voice.synthesize_stream_raw(text):
                int_data = np.frombuffer(audio_bytes, dtype=np.int16)
                ogg_file.write(int_data)

        return FileResponse(output_file, media_type="audio/ogg", filename=output_file)
    except Exception as e:
        return JSONResponse(content={"error": str(e)}, status_code=500)


@app.get("/")
def root():
    """
    Root endpoint to confirm the API is running.
    """
    return {"message": "Text-to-Speech API is up and running!"}