Intro

Trading buddy is a web client for trading with reliable, convenient and automatic position management and tracking

For more details check out article on my blog

Screenshots

Copied from the article actually

Docker Compose

Monitoring

Trading-Buddy features two monitoring stacks - appropriate one and overkill one:

• Appropriate one is Dozzle as a single container, you can see it under MONITORING > LIGHT section - it gathers and stores logs for all containers and shows real-time metrics for CPU and RAM usage and even features authentication. Eats at most 80Mi of RAM

• Overkill one is composed of:

  • Grafana Alloy for gathering metrics and logs
  • Victoria Metrics for storing metrics
  • Grafana Loki for storing logs
  • Grafana for displaying logs and metrics This stack eats more RAM than whole my app so there's not much point in utilizing it on VPS

CI/CD

For CI/CD both GitLab CI/CD and GitHub Actions could be used - both .yml files for each method are present in repo. GitLab is preffered option as sometimes deployment stage fails with GH actions due to some connections problem to remote virtual machine. Compose app is being launched on VM via exporting DOCKER_HOST and executing docker commands from runner on remote VM without need of copying sensitive data to the machine.

Kubernetes

In infra/kind-k8s you can find set of k8s resources which deploy Trading-Buddy into any k8s cluster

Monitoring

Monitoring stack is composed from components below, all of them need to be installed as helm charts:

• vm/victoria-metrics-k8s-stack - metrics gathering and storing
• grafana/loki - storage for logs
• grafana/k8s-monitoring - gathering of logs via Alloy and displaying metrics and logs via Grafana

GitOps

ArgoCD is used for monitoring changes in .yaml files in infra/kind-k8s

Running locally

Create all your .env files in the root of the project dir

.env.dev

Below is how your .env.dev must look like:

##### Frontend #####
VITE_API_BASE_URL=http://backend:8000/api/v1
VITE_API_SUFFIX=/api/v1
VITE_API_BE_BASE_URL=/api/v1
##### Backend #####
SECRET_KEY=<>
# True means launching backend from IDE, and database in docker, False - both backend and database in docker
DEBUG=True
ALLOWED_HOSTS=localhost,127.0.0.1,backend,nginx
CORS_CSRF_ALLOWED_ORIGINS=http://localhost,http://localhost:5173
SESSION_COOKIE_SECURE=False
CSRF_COOKIE_SECURE=False
CSRF_COOKIE_HTTPONLY=False
CSRF_USE_SESSIONS=False
CSRF_COOKIE_SAMESITE='Lax'
CORS_ALLOW_CREDENTIALS=True
##### Database #####
DB_NAME=postgres
# Name of db service in docker-compose.yml or name of the host (localhost when running on local machine and database has exposed port)
DB_HOST=localhost
DB_USER=<>
DB_PASSWORD=<>
##### SSO #####
GOOGLE_OAUTH_CLIENT_ID=<>
GOOGLE_OAUTH_SECRET=<>
LOGIN_REDIRECT_URL=/trading

SSO ones are optional - app will run fine without them as well

.env.prod

Production version is the same as dev one except for values below:

DEBUG=False
DB_HOST=db
ALLOWED_HOSTS=localhost,127.0.0.1,backend,nginx,<domain>
CORS_CSRF_ALLOWED_ORIGINS=http://localhost,http://localhost:5173,http://<domain>,https://domain

backup.env

Below is how your backup.env for dropbox backups must look like:

DROPBOX_REMOTE_PATH=<>
DROPBOX_APP_KEY=<>
DROPBOX_APP_SECRET=<>
DROPBOX_REFRESH_TOKEN=<>

If you want to run app ASAP you can skip backups setup, app will work fine without them as well.

Here is the guide I used to obtain all above variable values and setup dropbox for backing up in total: guide

Locally

Docker compose

If you just want to try out the project locally you can clone the repo and launch it in docker, for this you will need configured .env.dev

After creating and setting up .env.dev simpy run command below to launch app via docker from the root dir of the repo:

make compose-dev

or If you don't have make installed

docker compose -f docker-compose.dev.yml --env-file .env.dev up --build -d

IDE

For running from IDE you need to complete steps below:

1. Clone repo

   git clone https://github.com/avtorgenii/trading-buddy-2.0.git
   

2. Initialize venv for the backend:

   cd trading-buddy-2.0/trading_buddy_backend
   uv sync
   

   If you don't have uv yet here is the installation guide

3. Setup .env.dev as shown in section above

4. Initialize SvelteKit frontend as in README.md instruction in trading_buddy_fe dir:

   cd trading-buddy-2.0/trading_buddy_fe
   npm install
   

5. During development, I ran database only in docker container so just run command below from root of the project to create container for database:

   docker compose -f docker-compose.dev.yml --env-file .env.dev up -d db
   

Launch backend and frontend via CLI and app will be available at http://localhost:5173/:

# In backend dir - execute in separate shells
python manage.py runserver
python manage.py run_poller
python manage.py run_listeners


# In frontend dir
npm run dev

Every time you add changes to database scheme in models.py you have to run commands below for changes to be applied:

python manage.py makemigrations
python manage.py migrate

Dockerfile for backend applies migrations for you on the build stage

Usage

For more usage details check out article on my blog

Rules of usage

1. Currently restricted to open multiple trades on the same tool within one trading account, on app side you will be unable to do that, but if you open it on exchange side, handling orders and positions system will FAIL.
2. All decimals have maximum 20 digits, of which 12 are reserved for places after coma, which means that when BTC or any other trading tool will approach price of 100mln, code should be updated.
3. For now, manual cancellation of any order on the exchange side, breaks logic flow.
4. Exchange restricts placing Stop-loss order above current price, so unless your take-profits are meticulously near your entry price, or are take-losses actually, you should be fine, otherwise you may need to place stop-loss orders manually on exchange side
5. If you have opened position on exchange side, but it is not in database, it won't show up in your current and pending positions

General info

1. All datetimes are in UTC without any timezone info
2. Default mode of all trades is Cross Margin