Stock Analysis AI Agent

This tool uses Ollama with the deepseek-r1:7b model to analyze stocks and provide investment recommendations.

Features

• Technical analysis: Evaluates price movements, trends, and technical indicators
• Market analysis: Analyzes sector performance, market cap, and other market factors
• News analysis: Examines recent news related to the stock
• Final recommendation: Combines all analyses to provide an investment recommendation

Installation

For Mac Users

1. Make sure you have Python 3.8+ installed:

   python3 --version

2. Run the installation script (creates a virtual environment automatically):

   chmod +x install.sh
   ./install.sh

   This script will:

   • Create a Python virtual environment named venv
   • Activate the virtual environment
   • Install Ollama
   • Install all required Python packages

3. Start Ollama server in a separate terminal window:

   # In a separate terminal or tab:
   ollama serve

4. In another terminal window, pull the deepseek model:

   # In another terminal or tab:
   ollama pull deepseek-r1:7b

5. For future sessions, activate the virtual environment before running:

   source venv/bin/activate

Usage

1. Make sure your virtual environment is activated:

   source venv/bin/activate

2. Ensure the Ollama server is running in a separate terminal:

   # In a separate terminal window
   ollama serve

3. Run the analysis:

   python stock_analysis.py <STOCK_SYMBOL>

For example:

python stock_analysis.py AAPL    # Analyze Apple Inc.
python stock_analysis.py MSFT    # Analyze Microsoft
python stock_analysis.py GOOG    # Analyze Google/Alphabet

You can use country-specific suffixes for non-US stocks:

python stock_analysis.py INFY.NS  # Analyze Infosys (India)
python stock_analysis.py 9984.T   # Analyze SoftBank (Japan)

How It Works

The tool uses a workflow graph with the following nodes:

1. Technical Analysis: Calculates indicators like SMA, RSI, and volume trends
2. Market Analysis: Gathers information about sector, market cap, and PE ratio
3. News Analysis: Processes recent news about the stock
4. Recommendation: Combines all analyses to generate a final recommendation

The LLM (deepseek-r1:7b) processes each step and provides detailed analysis.

Requirements

• Python 3.8 or higher
• Ollama installed (https://ollama.com)
• Required Python packages:
  • langchain
  • langchain-core
  • langgraph
  • yfinance
  • langchain-openai
• An internet connection to fetch stock data

Troubleshooting for Mac

• If you get permission errors when running Ollama, you might need to run:

  sudo chmod +x /usr/local/bin/ollama

• If Ollama server exits with an error (Exit 1), try running it in a dedicated terminal window without the background process:

  # In a terminal dedicated to this purpose:
  ollama serve

• If you have issues with the virtual environment, you can create it manually:

  python3 -m venv venv
  source venv/bin/activate
  pip install langchain langchain-core langgraph yfinance langchain-openai

Example Output

The output provides a comprehensive analysis of the stock, including:

• Technical analysis with trend analysis and key signals
• Market analysis with sector performance and risk assessment
• News analysis with sentiment and potential impact
• Final recommendation with confidence score and target price range