Fixture for Augur tests

[Lilly-May]

PR Checklist

• Referenced issue is linked (see Rewrite Augur test with a fixture for sc_sim_augur #380)

Description of changes

Implements a fixture for the Augur tests as requested in issue #380.
Instead of loading the pre-calculated dataset sc_sim.h5ad that was stored in the tests/tools folder, the tests now use a fixture to load the sc_sim_augur dataset via the pertpy dataloader.
This only affects the tests for Augur and not the user directly.

Technical details

To ensure all tests still run (the results they expect are based on the pre-calculated sc_sim.h5ad dataset with 200 cells), the sc_sim_augur dataset (comprising 600 cells) is subsampled so that it contains the exact same 200 cells as the sc_sim.h5ad one (some brute-force testing might have been involved in finding the right random state for the subsampling...). That way, no changes in the actual tests were necessary, the fixture was simply added and used in all tests.

Performance-wise, there is no big difference (tested on my personal computer):

• Duration of all Augur tests before using a fixture: 15sec 313ms
• Duration of all Augur tests now (with fixture and dataset via dataloader): 13sec 677ms

[codecov]

Codecov Report

Merging #406 (3e0cbb0) into main (ea69f90) will not change coverage.
Report is 3 commits behind head on main.
The diff coverage is n/a.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #406   +/-   ##
=====================================
  Coverage   0.00%   0.00%           
=====================================
  Files         40      40           
  Lines       4877    4885    +8     
=====================================
- Misses      4877    4885    +8     

see 8 files with indirect coverage changes

[Zethson]

Great, thank you very much!