stock-backtesting-tool

Purpose

Created this tool to help backtest trading strategies while providing a frontend to be able to view backtesting outcomes.

Technologies/Libraries

• Python3
• FastAPI
• Ruff (code linting and formatting)
• Pandas (data manipulation)
• Poetry (dependency management)
• HTMX (frontend library)
• Docker

Code Structure

Within the app folder, a variation of the clean architecture was applied in order to have a separation of concerns between various files.

• clients: any kind of code that employs fetching data from a third-party data source would be (e.g., an API client that utilizes HTTP).
• models: domain object models that are used as a common language between different layers of the application.
• services: classes that might do some kind of data aggregation or data manipulation that might interact with different data sources.
• strategies: the stock trading execution strategies that determine when to buy and sell a stock.
• templates: the UI layer that is rendered in the FastAPI route.
• repositories (future work): classes that handle interacting with a data store of some kind.
• utils: code that handles generic general functions that could be used across different classes in the services folder.

Setup

This application uses .githooks to streamline development workflow using a library called pre-commit. Follow the instructions to install pre-commit itself.

pip install pre-commit

Once pre-commit is installed as a global dependency, run the following to install the .githooks:

pre-commit install

Ideas

• List out all of the strategies that you have currently running via an endpoint.
• Run a strategy
• Store data in a ClickHouse
  • ClickHouse, as a column-oriented OLAP DB, will allow for faster querying on larger datasets.

TODO List

• Refactor file generation code into different classes (i.e., )
• Dockerize application
• Add command line argument parsing to app/calculate.py CLI script
• Create Cron Job to Pull Down Stock Market Data
• Add Clickhouse to Store Stock Market Data
  • Store Hourly Stock Market Data
  • Store Stock Earning Dates

Architecture

NOTE: theoretical setup

flowchart LR
    frontend[HTMX Frontend]<-->FastAPI
    subgraph Render:
        direction LR
        FastAPI<-->yahoo-finance[Yahoo Finance]
    end

Loading