Feature Server

A price and feature detector written in Go 1.20.

Overview

Currently, the Feature Server subscribes to Binance's websocket feed for BTC/USDT on 5m candles. It calculates the following features and stores them in a MySQL database (PlanetScale):

• OHLC
• Volume
• Volume Moving Average

Issues, PRs, and general feedback are welcome!

How to contribute

Fork the repo (main branch) and submit a pull request.

TODO:

• Add more features (SMA, RSI, MACD, etc.)
• Add more symbols (ETH/USDT, XRP/USDT, etc.)
• Add a frontend to display the features!!

Self-host: Deploy to Google Cloud VM

To deploy on your own Google Cloud VM, follow the steps below.

Build the binary

docker build -t feature-server .

Tag the image and push to Google Artifact Registry (change the names to your own)

docker tag feature-server:latest asia.gcr.io/binance-websocket-feed-go/feature-server
docker push asia.gcr.io/binance-websocket-feed-go/feature-server

Create VM from image (if not already created)

gcloud compute instances create-with-container feature-server \
    --container-image asia.gcr.io/binance-websocket-feed-go/feature-server 

Update VM with new image (if already created)

gcloud compute instances update-container feature-server \
    --container-image asia.gcr.io/binance-websocket-feed-go/feature-server

Environment Setup

You'll need to have an .env file with:

ENV=development
DSN=<mysql://... REPLACE WITH YOUR OWN ...>

Then run the following commands to generate the Prisma client to the db package:

go run github.com/steebchen/prisma-client-go generate