openml
API can't filter datasets by uploader
Datasets can be listed with lots of filters, but strangely not by uploader user id.
I’d like to work on this issue. I will reproduce the problem and then propose a fix to allow filtering datasets by uploader user ID.
I can reproduce this issue. Can you confirm if the expected behavior is to support filtering by uploader ID through the API?
Thanks! Yes, but as the REST API is moving from PHP to Python let me bring in @PGijsbers to comment on this. If you want to fix this in the current REST API, that's also appreciated.
Thanks for the clarification! Since the API is moving from PHP to Python, I’d be happy to contribute to adding the uploader_id filter in the Python version. Could you please point me to the right repo or branch where the new Python REST API is being developed?
The development process for the new python server code isn't very streamlined nor well-documented yet, which doesn't really make this a beginner friendly issue for that API. I also haven't had time to work on it in a little while myself, which also makes it harder to give support :(
I think we should also evaluate how many filtering operations we should support through the traditional REST API. General search and filtering operations are perhaps better supported through Elastic Search, and we probably want to avoid our REST API endpoints from evolving to have feature-parity with elastic search.
For other OpenML beginner issues, we are currently recruiting some help in migrating our test suite in openml-python. It currently features tight integration testing (many calls to both the test and production servers) which we want less of, and also to move away from using the unittest framework to using pytest exclusively. It can be a good starting point for beginners, particularly the test framework conversion, since you're working on completely isolated (test) methods. PRs covering only one test at a time are preferred.
