How a Start Chat With take user Input Like Email

API
1

Hello, I just want to know a Chatgpt model takes the input from the user and start the after takes the input.

2

Hi @suresh24.enact - I hope I understand your question correctly. You can supply input as part of your prompt/user message (provided it does not exceed the maximum input token limit).

3

I want to using the api response then it will give me output according to that

4

I am still not entirely sure I understand the issue fully. Could you break down in greater detail what you would like to do? Thank you.

5

from flask import Flask, render_template, request, jsonify

from langchain.vectorstores import FAISS
from langchain.chat_models import ChatOpenAI
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.chains import ConversationalRetrievalChain
import os
from PyPDF2 import PdfReader

app = Flask(name)

Load PDF and extract text

pdf_file_obj = open(“docs/jat.pdf”, “rb”)
pdf_reader = PdfReader(pdf_file_obj)
num_pages = len(pdf_reader.pages)
detected_text = “”

for page_num in range(num_pages):
page_obj = pdf_reader.pages[page_num]
detected_text += page_obj.extract_text() + “\n\n”

pdf_file_obj.close()

Set OpenAI API Key

os.environ[“OPENAI_API_KEY”] = “”

Split text into documents

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
texts = text_splitter.create_documents([detected_text])

Build vector index

directory = “index_store”
vector_index = FAISS.from_documents(texts, OpenAIEmbeddings())
vector_index.save_local(directory)

Load vector index and set up retriever

vector_index = FAISS.load_local(directory, OpenAIEmbeddings())
retriever = vector_index.as_retriever(search_type=“similarity”, search_kwargs={“k”: 6})

Create ConversationalRetrievalChain

conv_interface = ConversationalRetrievalChain.from_llm(ChatOpenAI(temperature=0), retriever=retriever)

Store chat history

chat_history = # We won’t use this

@app.route(‘/’)
def index():
return render_template(‘chat.html’, user_input=“”, bot_response=“”)

@app.route(‘/api/chat’, methods=[‘POST’])
def chat():
user_input = request.form[‘user_input’]

# Get the bot's response
bot_response = conv_interface({"question": user_input, "chat_history": chat_history})["answer"]

return jsonify({"bot_response": bot_response})

if name == ‘main’:
app.run(debug=True, port=5001)

This code i need this give only my pdf knowledge base use only to give the answer