Headless algorithmic trading bot using market data from Alpaca API for real time data subscription/a paper trading account, and implemented using a bespoke Django wrapper for Alpaca's Python client.

Trades are placed with Alpaca Paper trading platform, with market data periodically sourced using Docker containers running Celery and Redis, and stored using a Postgresql database.

Current supported strategies:

• 14 day moving average
• 7 day moving average
• // TODO

• James Dalton

• Django - Django is a high-level Python Web framework that encourages rapid development and clean, pragmatic design.
• Django-Rest-Framework - Django REST framework is a powerful and flexible toolkit for building Web APIs.
• Docker - Docker is a set of platform as a service products that use OS-level virtualization to deliver software in packages called containers.
• Celery - Celery is a simple, flexible, and reliable distributed system to process vast amounts of messages, while providing operations with the tools required to maintain such a system. It’s a task queue with focus on real-time processing, while also supporting task scheduling.
• Redis - Redis is an open source (BSD licensed), in-memory data structure store, used as a database, cache and message broker
• Visual Studio Code - IDE

1. Install Docker Desktop. Create your own Alpaca API Account.

2. Navigate to your server directory, and create a .env file. Copy the included .env.example in this repo for a starter file;

$ cd server
$ cp .env.example .env

Note: setting POSTGRES_HOST_AUTH_METHOD=trust means Postgresql does not require a password. This is used only in development mode, being run on localhost

3. Activate your virtual environment, and navigate back into your project folder if necessary

$ pipenv shell

4. Run docker compose up to initialise django server, postgres for databasing, and celery/celery-beat/redis for handling of background tasks.

$ (server) docker compose up

Note: you may have to give wait-for-it.sh and docker-entrypoint.sh executable permissions

$ (server) chmod +x ./wait-for-it.sh && chmod +x ./docker-entrypoint.sh

5. To run the server for the first time, you will have to run database migrations. and create a superuser, and test data (refer core.management.bootstrap.py for detail on these commands). To do this you can use docker compose and the web container

# Run migrations
$ (server) docker compose exec web python manage.py migrate
# Create superuser account
$ (server) docker compose exec web python manage.py bootstrap -cs
# Create example data
$ (server) docker compose exec web python manage.py bootstrap -ced

6. Create a strategy via POST to http://localhost:8000/v1/strategies/ or through your browser using Django's admin at http://localhost:8000/admin/core/strategy/

7. Keep docker and your local server running for periodic tasks to perform the given strategy created in 6.

$ (server) docker compose exec web python manage.py test

• Optionally add --keepdb to persist database between tests, and --verbosity=2 to receive verbose output of tests

$ (server) docker compose exec web python manage.py test --keepdb --verbosity=2

[packages]
django
djangorestframework
django-environ
alpaca-trade-api
celery
django-celery-beat
psycopg2
psycopg2-binary

[dev-packages]
django-debug-toolbar
freezegun

To make running docker compose commands on your local machine easier, use alias' in .bashrc/.zshrc configuration files eg.

# ~/.zshrc file
[...]
alias pm="docker compose exec web python manage.py"

Then the command docker compose exec web python manage.py test server becomes pm test server

This project is licensed under the terms of the MIT license