import gradio as gr
import numpy as np
import runpod
import os

RUNPOD_API = os.environ.get("RUNPOD_API")
RUNPOD_ENDPOINT = os.environ.get("RUNPOD_ENDPOINT")


def process_transcribe(file):
    audio_nparray = file[1]
    audio_list = audio_nparray.tolist()
    
    runpod.api_key = RUNPOD_API
    endpoint = runpod.Endpoint(RUNPOD_ENDPOINT)

    run_request = endpoint.run_sync(
        {"audio_list": audio_list}
    )
    raw_text = run_request
    
    return raw_text

with gr.Blocks(theme=gr.themes.Default(primary_hue="blue")) as iface:
    with gr.Tab("Speech2text prescription"):
        with gr.Row():
            with gr.Column():
                audio_file = gr.Audio(sources=["upload", "microphone"], type="numpy", label="Audio")
                submit_btn = gr.Button("Submit", variant="primary")
            with gr.Column():
                raw_transcript = gr.Textbox(label="Transcription")
        submit_btn.click(process_transcribe, inputs=[audio_file], outputs=[raw_transcript])
iface.launch(debug=True)