AutoTrader is Python-based platform intended to help in the development, optimisation and deployment of automated trading systems. A basic level of experience with Python is recommended for using AutoTrader, but the docs aim to make using it as easy as possible with detailed tutorials and documentation.
- Version 0.7 has been released, adding integrations with CCXT and dYdX crypto exchanges. Many more powerful upgrades too.
- AutoTrader has been featured in GitClone's recent article, Top Crypto Trader Open-Source Projects on Github.
- A feature-rich trading simulator, supporting backtesting and papertrading. The 'virtual broker' allows you to test your strategies in a risk-free, simulated environment before going live. Capable of simulating multiple order types, stop-losses and take-profits, cross-exchange arbitrage and portfolio strategies, AutoTrader has more than enough to build a profitable trading system.
- Integrated data feeds, making OHLC data retrieval as easy as possible.
- Automated interactive visualisation using Bokeh
- Library of custom indicators.
- Live trading supported for multiple venues.
- Detailed documenation and tutorials
- Repository of example strategies
| Broker | Asset classes | Integration status |
|---|---|---|
| Oanda | Forex CFDs | Complete |
| Interactive Brokers | Many | In progress |
| dYdX | Cryptocurrencies | Complete |
| CCXT | Cryptocurrencies | In progress |
AutoTrader can be installed using pip:
pip install autotrader
AutoTrader can be updated by appending the --upgrade flag to the install command:
pip install autotrader --upgrade
AutoTrader is very well documented in-code and on Read the Docs. There is also a detailed walthrough, covering everything from strategy concept to livetrading.
Example strategies can be found in the demo repository.
The chart below is produced by a backtest of the MACD trend strategy documented in the tutorials (and available in the demo repository). Entry signals are defined by MACD crossovers, with exit targets defined by a 1.5 risk-to-reward ratio. Stop-losses are automatically placed using the custom swing detection indicator, and position sizes are dynamically calculated based on risk percentages defined in the strategy configuration.
Running this strategy with AutoTrader in backtest mode will produce the following interactive chart.
Note that stop loss and take profit levels are shown for each trade taken. This allows you to see how effective your exit strategy is - are you being stopped out too early by placing your stop losses too tight? Are you missing out on otherwise profitable trades becuase your take profits are too far away? AutoTrader helps you visualise your strategy and answer these questions.
To contribute to autotrader, please read the instructions below,
and stick to the styling of the code.
-
Create a new Python virtual environment to isolate the package. You can do so using
venvor anaconda. -
Install the code in editable mode using the command below (run from inside the
autotraderroot directory). Also install all dependencies using the[all]command, which includes the developer dependencies.
pip install -e .[all]
- Install the pre-commit hooks.
pre-commit install
- Start developing! After following the steps above, you are ready to start developing the code. Make sure to follow the guidelines below.
-
Fork the repository and clone to your local machine for development.
-
Run black on any code you modify. This formats it according to PEP8 standards.
-
Document as you go: use numpy style docstrings, and add to the docs where relevant.
-
Write unit tests for the code you add, and include them in
tests/. This project uses pytest. -
Commit code regularly to avoid large commits with many changes.
-
Write meaningful commit messages, following the Conventional Commits standard. The python package commitizen is a great tool to help with this, and is already configured for this repo. Simply stage changed code, then use the
cz ccommand to make a commit. -
Open a Pull Request when your code is complete and ready to be merged.
To build the documentation, run the commands below.
cd docs/
make html
xdg-open build/html/index.html
If you are actively developing the docs, consider using
sphinx-autobuild.
This will continuosly update the docs for you to see any changes
live, rather than re-building repeatadly. From the docs/
directory, run the following command:
sphinx-autobuild source/ build/html --open-browser
AutoTrader is licensed under the GNU General Public License v3.0.
This platform is currently under heavy development and should not be considered stable for livetrading until version 1.0.0 is released.
Never risk money you cannot afford to lose. Always test your strategies on a paper trading account before taking it live.