app.py

import os
import gradio as gr
import openai

from langdetect import detect 
from gtts import gTTS
from pdfminer.high_level import extract_text

openai.api_key = os.environ['OPENAI_API_KEY']

user_db = {os.environ['username1']: os.environ['password1'], os.environ['username2']: os.environ['password2']}

messages = [{"role": "system", "content": 'You are a helpful assistant.'}]



def roleChoice(role):
    global messages
    messages = [{"role": "system", "content": role}]
    return "role:" + role


def audioGPT(audio):
    global messages 

    audio_file = open(audio, "rb")
    transcript = openai.Audio.transcribe("whisper-1", audio_file)

    messages.append({"role": "user", "content": transcript["text"]})

    response = openai.ChatCompletion.create(model="gpt-4", messages=messages)

    system_message = response["choices"][0]["message"]
    messages.append(system_message)

    chats = ""
    for msg in messages:
        if msg['role'] != 'system':
            chats += msg['role'] + ": " + msg['content'] + "\n\n"

    return chats


def textGPT(text):
    global messages

    messages.append({"role": "user", "content": text})

    response = openai.ChatCompletion.create(model="gpt-4", messages=messages)

    system_message = response["choices"][0]["message"]
    messages.append(system_message)

    chats = ""
    for msg in messages:
        if msg['role'] != 'system':
            chats += msg['role'] + ": " + msg['content'] + "\n\n"

    return chats


def siriGPT(audio):
    global messages 

    audio_file = open(audio, "rb")
    transcript = openai.Audio.transcribe("whisper-1", audio_file)

    messages.append({"role": "user", "content": transcript["text"]})

    response = openai.ChatCompletion.create(model="gpt-4", messages=messages)

    system_message = response["choices"][0]["message"]
    messages.append(system_message)

    lang = detect(system_message['content'])

    narrate_ans = gTTS(text=system_message['content'], lang=lang, slow=False) 
    narrate_ans.save("narrate.wav") 

    return "narrate.wav"
    

def fileGPT(prompt, file_obj):
    global messages 

    file_text = extract_text(file_obj.name)
    text = prompt + "\n\n" + file_text
    
    messages.append({"role": "user", "content": text})
    
    response = openai.ChatCompletion.create(model="gpt-4", messages=messages)

    system_message = response["choices"][0]["message"]
    messages.append(system_message)

    chats = ""
    for msg in messages:
        if msg['role'] != 'system':
            chats += msg['role'] + ": " + msg['content'] + "\n\n"

    return chats



def clear():
    global messages
    messages = [{"role": "system", "content": 'You are a helpful technology assistant.'}]
    return
    
def show():
    global messages
    chats = ""
    for msg in messages:
        if msg['role'] != 'system':
            chats += msg['role'] + ": " + msg['content'] + "\n\n"

    return chats


with gr.Blocks() as chatHistory:
    gr.Markdown("Click the Clear button below to remove all the chat history.")
    clear_btn = gr.Button("Clear")
    clear_btn.click(fn=clear, inputs=None, outputs=None, queue=False)

    gr.Markdown("Click the Display button below to show all the chat history.")
    show_out = gr.Textbox()
    show_btn = gr.Button("Display")
    show_btn.click(fn=show, inputs=None, outputs=show_out, queue=False)



role = gr.Interface(fn=roleChoice, inputs="text", outputs="text", description = "Choose your GPT roles, e.g. You are a helpful technology assistant. 你是一位 IT 架构师。 你是一位开发者关系顾问。你是一位机器学习工程师。你是一位高级 C++ 开发人员 ")
text = gr.Interface(fn=textGPT, inputs="text", outputs="text")
audio = gr.Interface(fn=audioGPT, inputs=gr.Audio(source="microphone", type="filepath"), outputs="text")
siri = gr.Interface(fn=siriGPT, inputs=gr.Audio(source="microphone", type="filepath"), outputs = "audio")
file = gr.Interface(fn=fileGPT, inputs=["text", "file"], outputs="text", description = "Enter prompt sentences and your PDF. e.g. lets think step by step, summarize this following text:  或者 让我们一步一步地思考，总结以下的内容:")
demo = gr.TabbedInterface([role, text, audio, siri, file, chatHistory], [ "roleChoice", "chatGPT", "audioGPT", "siriGPT", "fileGPT", "ChatHistory"])

if __name__ == "__main__":
    demo.launch(enable_queue=False, auth=lambda u, p: user_db.get(u) == p,
        auth_message="Welcome to Yichuan GPT!")
    #demo.launch()