Page MenuHomePhabricator

Split WDQS git repository
Closed, ResolvedPublic

Description

Currently all code related to the WDQS resides in one big git repository (https://gerrit.wikimedia.org/r/q/project:wikidata/query/rdf). As part of the migration to a new triple store backend several components in this repo need to be rewritten, ported or even dropped.

Going forward this repository is supposed to be split into several new repositories with the old one being archived. This split will most likely not happen in one go but in several steps, each migrating one component to a new repository.
This task is intended to keep track of this process.

AC:

  • A decision has been taken regarding the home of new repositories (GitLab vs. Gerrit)
  • All components (which are not deprecated after the backend transition) have been migrated to a new repo
  • The original repo has been archived (marked read-only)

Event Timeline

trueg updated the task description. (Show Details)

The split takes place in naturally occuring steps during the migration to the new WDQS backend. The wdqs-proxy and the wdqs-mwsparql services have already been split into their own git repositories.

trueg changed the task status from Open to In Progress.Mar 26 2026, 1:45 PM
trueg claimed this task.
trueg updated the task description. (Show Details)
trueg updated the task description. (Show Details)
DSantamaria changed the status of subtask T422054: wdqs-streaming-consumer: Stand-alone git repository from Open to In Progress.

Current status:

The following components have been separated into their own repositories:

Open Questions:

  • The original repository contains a set of command line tools which have not been transitioned into their own Gitlab repositories. It is unclear which are still required and which are deprecated. This will become more clear as the migration effort to the new triple store continues.

@trueg, what is the current status of this one?

@DSantamaria the status is unchanged. We have separated repositories as listed above and the question regarding the tools is still open. I would like to keep this open to eventually document the decision on the remaining code in tools which is not used in the post-migration plans yet.

trueg updated the task description. (Show Details)

After some discussion within the team, I will resolve this issue as being completed. While we did not extract all code from the original rdf repository yet, we did create all the new Gitlab repositories we need. Any additional tools can be merged into one of those new repositories, if required. They will most likely be debugging tools.