ckan / datapusher

A standalone web service that pushes data files from a CKAN site resources into its DataStore
GNU Affero General Public License v3.0
77 stars 154 forks source link

ref: Add Dockerfile and Github Actions pipeline #234

Open gustavovalverde opened 3 years ago

gustavovalverde commented 3 years ago

Fixes #110 Fixes #136 Fixes #163 Closes #210

gustavovalverde commented 3 years ago

cc: @kowh-ai this is using a very similar approach to https://github.com/ckan/ckan-docker/blob/master/datapusher/Dockerfile

But it's also adding the whole workflow for Github Actions to deploy the images when pushing to master and also when creating a release.

I'm not sure if your team is focusing on using Github Registry or DockerHub, but both are configured in the pipeline.

mbocevski commented 3 years ago

@gustavovalverde check out https://github.com/keitaroinc/docker-ckan for CKAN, datapusher docker images.

gustavovalverde commented 3 years ago

Thanks @mbocevski

I've seen a few repos that are already building the images, aside from this one you're sharing. But I'm not sure if this is something that's going to be finally merged on each CKAN source code repo: datapusher, CKAN, solr, etc; or if this images are going to be built and pushed to CKAN official accounts.

Just to be sure if there's an official roadmap or agreement on how those images are going to be handled; as we're going to deploy this suite on my country government, but we're required to use official sources, that's why I started this PR here.

One of my mains concerns is also seeing the use of git sources inside the Dockerfile https://github.com/keitaroinc/docker-ckan/blob/master/images/datapusher/Dockerfile#L10 as this makes me wonder if the Dockerfile won't reside on the original source code anymore.

mbocevski commented 3 years ago

@gustavovalverde Keitaro has been publishing and maintaining CKAN and CKAN related docker images for over 5 years, we have over 10k docker image pulls for our CKAN docker images. We do use our fork of datapusher due to changes that have not been merged upstream due to datapusher not getting a lot of maintenance time from the core team.

My suggestion is to use Keitaro's docker images and feel free to log issues on our repo if you run into any container related issues. We also have a Helm chart for Linux https://github.com/keitaroinc/ckan-helm

gustavovalverde commented 3 years ago

Thanks for your answer @mbocevski

If you accept PRs on your side, I could also create some issues with their related PRs for the Dockerfiles and pipeline.

mbocevski commented 3 years ago

Yeah sure