Alpaca Trading Simulator

This project is a multi-portfolio trading simulator built with Python. It uses Alpaca's API along with other data sources and OpenAI for strategy generation.

Setup

1. Create a .env file based on .env.example and fill in your API keys.

2. Install requirements:

   pip install -r requirements.txt

3. Run the environment test:

   python -m tests.env_test

4. Test the portfolio manager (requires Alpaca paper account keys):

   python -m tests.portfolio_test

5. Run the research engine test (requires network and API keys for best results):

   python -m tests.research_test

6. Try the OpenAI strategy demo:

   python -m tests.strategy_test

7. Run the risk management test:

   python -m tests.risk_test

8. Run the diversification analysis test:

   python -m tests.diversification_test

9. Generate an example trade export and report:

   python -m tests.report_test

10. Run the Flask dashboard:

python app.py

The dashboard lets you pick a strategy for each portfolio from a dropdown. Each portfolio requires its own Alpaca API key and secret. The create form in the dashboard asks for these credentials – one key pair cannot be shared between multiple portfolios. The "Step" button accepts an optional comma- separated list of symbols. If left blank, the simulator will trade a set of trending tickers automatically. Set TRENDING_SOURCE=openai in your .env file to let ChatGPT suggest trending symbols instead of using Yahoo Finance. The Buy Only button works similarly but ignores sell signals and only executes trades when a buy recommendation is returned. Select debug in the Activity filter to view the exact prompts sent to OpenAI and the received responses. A small form inside each portfolio block also lets you submit manual market orders. Provide a symbol, quantity and choose buy or sell. Manual trades are stored in the history marked with [M].

Custom Prompt Editor

Within the dashboard each portfolio provides a textarea to configure a custom prompt for OpenAI. The prompt must contain the placeholders {strategy_type}, {portfolio} and {research} which will automatically be filled in before the request is sent. Use the "Preview" button to see the generated decision before saving the prompt.