main.py

import os
import openai
from flask import Flask, request, jsonify, send_file
from transformers import pipeline
from openai import OpenAI
from gtts import gTTS

app = Flask(__name__)
client = OpenAI()

openai.api_key = os.getenv("OPENAI_API_KEY")
pipe = pipeline(model="seeafricatz/kiaziboraasr")

def transcribe(audio_path):
    text = pipe(audio_path)["text"]
    return text

def generate_response(transcribed_text):
    response = client.chat_completions.create(
        messages=[
            {"role": "system", "content": "All your answers should be in Swahili only..."},
            {"role": "user", "content": transcribed_text},
        ],
        model="gpt-4"
    )
    return response.choices[0].message.content

def inference(text):
    tts = gTTS(text, lang='sw')
    output_file = "tts_output.mp3"
    tts.save(output_file)
    return output_file

@app.route('/process_audio', methods=['POST'])
def process_audio():
    if 'audio' not in request.files:
        return jsonify({'error': 'No audio file provided'}), 400

    audio_file = request.files['audio']
    audio_path = "temp_audio.wav"
    audio_file.save(audio_path)

    transcribed_text = transcribe(audio_path)
    response_text = generate_response(transcribed_text)
    output_file = inference(response_text)

    return jsonify({
        'response_text': response_text,
        'response_audio_url': f'/audio/{output_file}'
    })

@app.route('/audio/<path:filename>')
def audio(filename):
    return send_file(filename, as_attachment=True)

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=8000)