Build your worker image¶
Local docker build¶
To build a docker image of the worker we've previously built, we can use the
datashare-python script.
It's a tiny wrapper around the docker compose CLI:
./datashare-python build datashare-python=> [datashare-python internal] booting buildkit 0.9s
=> => starting container buildx_buildkit_strange_curran0 0.9s
=> [datashare-python internal] load build definition from Dockerfile
...
=> => exporting layers 56.1s
=> => exporting manifest sha256:7462a3f43df6073c57fc2482726a65d43e4f83f68ccd098ec0804b8b959d9a17 0.0s
=> => exporting config sha256:184ed641f1ee82d4eb068143702d3cbec32b25413051764000353b02458e12a1 0.0s
=> => sending tarball 38.7s
=> [datashare-python datashare-python] importing to docker
=> => starting container buildx_buildkit_strange_curran0 0.9s
=> [datashare-python internal] load build definition from Dockerfile
...
=> => exporting layers 56.1s
=> => exporting manifest sha256:7462a3f43df6073c57fc2482726a65d43e4f83f68ccd098ec0804b8b959d9a17 0.0s
=> => exporting config sha256:184ed641f1ee82d4eb068143702d3cbec32b25413051764000353b02458e12a1 0.0s
=> => sending tarball 38.7s
=> [datashare-python datashare-python] importing to docker
Publish to Docker Hub¶
You can also fork the template repository and use the CI to build and publish the worker image to Docker Hub.
To publish, make sure to set the DOCKERHUB_USERNAME and DOCKERHUB_TOKEN (see documentation)
and then just create a tag to trigger the build.