[BUG-python/deployment] filter_by returning unexpected results for response_status

[bertozzivill]

Describe the bug
I'm working with a RemoteFeedbackDataset from Argilla, and I'm trying to filter by response status, but the filtering system seems to be returning false positives.
As an example, filtering down only to a reference called "quinones1997anopheles" and a record type called "EntomologicalOutcome" returns a dataset with a single record, with one response whose status is "submitted". Filtering on the same values, but adding a response_status filter for "submitted" returns that same record, as expected. However, filtering on the same values while adding a response_status filter for "draft" also returns that same record. I would expect it to return an empty dataset since the filtration shouldn't have captured anything, right? Thanks for your help.

Stacktrace and Code to create the bug
See image above, I apologize that I don't have an easily replicable example but am hoping you could test this on an internal example to recreate the bug.

Expected behavior
I expected to only receive records with at least one "draft" response when using filter_by(response_status=["draft"])

Environment:
│ • Client version: 0.1.0a3 │
│ • Server version: 0.1.0a4 │
│ • ElasticSearch version: 8.5.1

[davidberenstein1957]

hi @bertozzivill would you be able to provide us with the version you are using?

[bertozzivill]

oh sorry, I thought that was captured by the output of python -m argilla info. I'm using argilla version 1.29.0

[davidberenstein1957]

Hi @bertozzivill , thank you for the update. As of now, we don't plan to create a patch for the 1.29 version in the short term, however, you could move over to the 2.0 version of Argilla to get more extensive bug fixes and support.

[bertozzivill]

thanks @davidberenstein1957. Unfortunately I can't update at the moment since the platform I'm using was built with a dependency on version 1.x. I'll try to find a solution that doesn't rely on filter_by().