| .vscode | ||
| cron | ||
| requirements | ||
| xm | ||
| .dockerignore | ||
| .gitignore | ||
| .pre-commit-config.yaml | ||
| __init__.py | ||
| bot.py | ||
| config.sample.py | ||
| docker-compose.debug.yml | ||
| docker-compose.yml | ||
| Dockerfile | ||
| fxstudy.ipynb | ||
| logging.conf | ||
| Pipfile | ||
| Pipfile.lock | ||
| pyproject.toml | ||
| README.md | ||
| requirements.txt | ||
| xmglobal5setup.exe | ||
Trading ML Experiment
This is a study on how different Neural Networks would be able to find the right way of modelling and finding a pattern through limited past or backtests before attempting to place a trade.
Develop in Python 🐍 with all of the packages mentioned in the requirements.txt. Primarily using Pytorch to model a Neural Network.
Actually, Pytorch is a bit advanced for a noob like me. So I used sci-kit learn and just get a working model with the defaults first.
Usage
Other than the required pip to install, the Metatrader client must have "Autotrading" mode enabled.
Copy and paste the config.sample.py file to config.py and change the contents to fit for your account and pairs that you want to trade.
Then run,
python3 app.py
Development and Contribution
This repo utilises pre-commit and manages package management with pipenv.
To set up the dev environment, pipenv install -d and pre-commit install.
TODO
- [] Get a working docker container of MT5 running
- [] Have the local
app.pyfile point to the running MT5 docker container instance - [] Containerize
app.pywhile pointing to the running MT5 docker container instance - [] Add cron job to execute
app.pycontainer while pointing to the running MT5 docker container instance