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

#any vector server should work, trying pinecone first
import pinecone

#langchain part
from langchain.llms import OpenAI
from langchain.text_splitter import SpacyTextSplitter
from langchain.document_loaders import TextLoader
from langchain.document_loaders import DirectoryLoader
from langchain.indexes import VectorstoreIndexCreator
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.vectorstores import Pinecone


openai.api_key = os.environ['OPENAI_API_KEY']
pinecone_key = os.environ['PINECONE_API_KEY']
pinecone_environment='us-west1-gcp-free'


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

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


#debug use
def foo(dir):
    return [d.name for d in dir]








def init_pinecone(index_name):
    # initialize connection to Pinecone vector DB (app.pinecone.io for API key)
    pinecone.init(
        api_key=pinecone_key,
        environment=pinecone_environment
    )
    #using openai embedding hence dim = 1536
    pinecone.create_index(index_name, dimension=1536, metric="euclidean")
    index = pinecone.Index(index_name)
    return index


def process_file(index_name, docs):
    index = init_pinecone(index_name)
    embeddings = OpenAIEmbeddings(openai_api_key=os.environ['OPENAI_API_KEY'])
    
    loader = DirectoryLoader(docs.name, glob="*.txt", loader_cls=TextLoader)
    documents = loader.load()
    #print(documents)


    #pipeline='zh_core_web_sm'
    splter = SpacyTextSplitter(chunk_size=1000,chunk_overlap=200)
    split_text = splter.split_documents(documents)
    
    for document in split_text:
    	Pinecone.from_documents([document], embeddings, index_name=index_name)

    return list_pinecone(index_name)


def list_pinecone(index_name):
    index = pinecone.Index(index_name)
    stats = index.describe_index_stats()
    return stats





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






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 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)


# debug use
with gr.Blocks() as foo:
    input = gr.File(file_count="directory")
    files = gr.Textbox()
    show = gr.Button(value="Show")
    show.click(foo, input, files)

    



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")

pinecone = gr.Interface(fn=process_file, inputs=["text", gr.inputs.File(file_count="directory")], 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, file, chatHistory, pinecone, foo], [ "roleChoice", "chatGPT", "fileGPT", "ChatHistory", "Pinecone", "foo"])

if __name__ == "__main__":
    demo.launch(enable_queue=False, auth=lambda u, p: user_db.get(u) == p,
        auth_message="This is not designed to be used publicly as it links to a personal openAI API. However, you can copy my code and create your own multi-functional ChatGPT with your unique ID and password by utilizing the 'Repository secrets' feature in huggingface.")
    #demo.launch()