Finance-NLP

Overview

Finance-NLP is a comprehensive project that leverages Natural Language Processing (NLP) to analyze and extract valuable insights from financial documents such as 10-K filings. The project includes an AI model for processing financial data, a FastAPI backend for serving the model, and a React-based frontend for interacting with the chatbot.

Project Structure

• AI Model: The AI model is built using transformers and is designed to process financial documents, extract relevant information, and answer questions based on the content.
• Backend: The backend is implemented using FastAPI, which serves the AI model and handles API requests from the frontend.
• Frontend: The frontend is a React application that provides a user-friendly interface for interacting with the chatbot.

AI Model

The AI model is based on GPT-2 and fine-tuned on financial data to understand and respond to queries related to financial documents. The model is loaded and served using FastAPI.

Key Components

• Tokenizer and Model: The GPT-2 tokenizer and model are loaded using the transformers library.
• Chatbot Response Function: A function to generate responses based on user input.
• API Endpoint: A FastAPI endpoint to handle POST requests and return the chatbot's response.

Backend

The backend is implemented using FastAPI and includes the following key components:

• CORS Middleware: Configured to allow cross-origin requests from the frontend.
• API Endpoint: A POST endpoint /chat/ that receives user input, processes it using the AI model, and returns the response.

Example Code

from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
import uvicorn

app = FastAPI()

app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

model_name = "gpt2"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained("chatbot_model/checkpoint-3499")
tokenizer.pad_token = tokenizer.eos_token

class ChatRequest(BaseModel):
    user_input: str

def chatbot_response(user_input):
    input_text = f"User: {user_input}\nBot:"
    inputs = tokenizer(input_text, return_tensors="pt")
    input_ids = inputs.input_ids
    attention_mask = inputs.attention_mask

    with torch.no_grad():
        output = model.generate(input_ids, attention_mask=attention_mask, max_length=50, pad_token_id=tokenizer.eos_token_id)

    response = tokenizer.decode(output[:, input_ids.shape[-1]:][0], skip_special_tokens=True)
    return response.strip()

@app.post("/chat/")
def chat(request: ChatRequest):
    response = chatbot_response(request.user_input)
    return {"bot_response": response}

if __name__ == "__main__":
    uvicorn.run(app, host="127.0.0.1", port=8000)