📈 Stock Ticker Analysis Application

This application is designed to provide a comprehensive analysis of a given stock ticker. The workflow involves checking the current price of the ticker, searching for relevant articles, filtering for the most relevant articles, and generating a summary report with sentiment analysis. The application is built using FastAPI and Ngrok to run and expose the API endpoint.

llm-stock-demo.mp4

🌟 Table of Contents

• 📈 Stock Ticker Analysis Application
  • 🌟 Table of Contents
  • 🚀 Installation
    • LLM API
    • Master Service
    • Frontend
  • 🛠️ Architecture
  • 🔍 API Services
    • LLM Service
      • Usage
    • Analysis/Master Service
      • Usage
  • ⚙️ Workflow
  • 📃 License

🚀 Installation

LLM API

This is a self hosted LLM model locally using Ollama. The model I will be using is llama3

Prerequisite: Ollama installed on your machine

ollama run llama3

Step 1: Install dependencies

cd llm_service
pip install -r requirements.txt

Step 2: Run scripts to start the fastapi server:

Linux

./script.sh

Window

./script.bat

• Server will run on localhost:8001

Master Service

This is the main API for the stock analysis utilizing search engine Tavily and our own LLM API in LLM service

Configure environment variable .env.example

TAVILY_API_KEY=""
STOCK_API_KEY=""
LLM_SERVICE_URL=http://127.0.0.1:8001

Visit twelveapi for the STOCK_API_KEY

Visit Tavily for TAVILY_API_KEY

Step 1: Install dependencies

cd master_agent
pip install -r requirements.txt

Step 2: Run scripts to start the fastapi server:

Linux

./script.sh

Window

./script.bat

• Server will run on localhost:8000

Frontend

I use streamlit for a quick prototype for this application. Step 1: Install dependencies

cd frontend
pip install -r requirements.txt

Step 2: Run scripts to start the fastapi server:

Linux

./script.sh

Window

./script.bat

🛠️ Architecture

1. Price Agent: Use Twelve API to retrieve the current stock price

2. Search: Search for relevant articles for analysis

3. Filter Agent: Filter the article by score based on search results

4. LLM Agent: Use RAG technique with self-hosted llama3 model for sentiment analysis

   

🔍 API Services

LLM Service

The Text Analysis API allows users to send a text message and receive a generated analysis based on the llama3 model from Ollama.

Usage

curl -X POST "http://127.0.0.1:8001/generate" -H "Content-Type: application/json" -d '{"content": "Analyze this text"}'

import requests
url = "http://127.0.0.1:8001/generate"
res = requests.post(url, json={"content": "Hi, how are you today?"})

Analysis/Master Service

Run the workflow to retrieve the stock analysis utilizing the llama3 service and RAG technique

Usage

curl -X 'GET' \
  'http://example.com/ticker?ticker=AAPL' \
  -H 'accept: application/json'

import requests

url = 'http://127.0.0.1:8000/ticker'
params = {'ticker': 'AAPL'}
headers = {'accept': 'application/json'}

response = requests.get(url, params=params, headers=headers)
data = response.json()

print(data)

⚙️ Workflow

1. User Input:

   • The user provides a stock ticker.

2. Price Agent:

   • The application checks the current price of the ticker. If the ticker does not exist, the application returns an error and ends the process.

3. Search Agent:

   • If the ticker exists, the application searches for articles and sources related to the stock ticker.

4. Article Filtering:

   • The articles are filtered based on the score from Tavily search tool

5. Summary Report:

   • The most relevant article is then passed into the LLM again to write a summary report and sentiment analysis.
   • The summary report is returned to the user.

📃 License

This project is licensed under the MIT License.