Build custom worker images with docker

Requirements

• Docker installed on your machine

Create your Dockerfile

Create a Dockerfile in your project directory:

ARG VERSION=latest
 
FROM ghcr.io/arcadeai/worker-base:${VERSION}
 
# Copy the file that lists all your desired toolkits
COPY toolkits.txt ./
 
# Install these toolkits
RUN pip install -r toolkits.txt

With this Dockerfile, we start from the Arcade worker base image, copy in a file named toolkits.txt, and install each toolkit listed there.

List Your Toolkits

Create a file named toolkits.txt in the same directory. Add the toolkits you want installed:

arcade-google
arcade-web
arcade-zoom

Adjust this file as needed. Each line in toolkits.txt should specify a Python package name and version.

3. Build the Image

From the directory containing the Dockerfile and toolkits.txt, run:

docker build -t custom-worker:0.1.0 .

This command creates a Docker image named custom-worker with the tag 0.1.0 using the instructions in your Dockerfile.

Run the Worker

To start the worker container:

docker run -p 8002:8002 \
  -e ARCADE_WORKER_SECRET="your_secret_here" \
  custom-worker:0.1.0

Replace “your_secret_here” with a secret of your choice. Your engine will need access to this secret to call your worker. The worker will be accessible on port 8002 of your local machine.