import requests
import gradio as gr
import base64

speakers = []
with open('speakers.txt', 'r') as fp:
    for line in fp:
        speakers.append(line.strip())

def synthesize(key, sentence, speaker):
    headers = {
      'Authorization': 'Bearer {}'.format(key),
      'Content-Type': 'application/json',
    }
    
    json_data = {
      'text': sentence,
      'speaker': speaker,
    }
    
    response = requests.post('https://rjmopratfrdjgmfmaios.functions.supabase.co/rime-tts', headers=headers, json=json_data)
    audioContent = response.json()['audioContent']
    decode_string = base64.b64decode(audioContent)
    with open('tmp.wav', 'wb') as fp:
        fp.write(decode_string)
    return 'tmp.wav'

iface = gr.Interface(fn=synthesize, inputs=[gr.Textbox(type="password", info="This is where you put your Rime TTS API key."), 
                                            gr.Textbox(info="Enter the text you want synthesized here. Check out our documentation for specifics and suggestions on text input!"),
                                           gr.Dropdown(list(speakers), info="Enter the code for the speaker you want to synthesize. (see docs/voices/)")], outputs=gr.Audio())
iface.launch()