Database error - on upload of a run from OpenML-Python

Description

An OpenMLServerException occurs for a particular unit test which executes a scikit-learn model on a task and tries to publish. The raised exception:
openml.exceptions.OpenMLServerException: Database error. Setup search query failed - None

A post request is made which receives a 412 response and therefore an exception is raised during the check where the status code is not 200.

Steps/Code to Reproduce

Given the unit test invokes a series of locally defined functions, one simple way to run this single unit test could be to clone the OpenML-Python repo and run:
pytest -k 'test_run_and_upload_decision_tree_pipeline

Given that the execution stack for the unit test case appears to run without errors, the actual post request made during publish can be found here.
The data and file_elements packaged in the call are attached as pickle files for reference, while the URL used for the post request is: https://test.openml.org/api/v1/xml/run/

Run pytest -k 'test_run_and_upload_decision_tree_pipeline' from the OpenML-Python repository and inspect the publish request in openml/_api_calls.py at line 207, along with tests/test_runs/test_run_functions.py at line 570. Use the attached pickle files and test.openml.org endpoint details to investigate the 412 response; done means the upload no longer raises the database error and the test passes.