Speedup For EDistance-like distances

[selmanozleyen]

No description provided.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 66.91729% with 44 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.83%. Comparing base (12897e1) to head (98f3c64).
⚠️ Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
pertpy/tools/_distances/_distances.py	66.91%	44 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #880      +/-   ##
==========================================
- Coverage   73.54%   71.83%   -1.71%     
==========================================
  Files          48       48              
  Lines        5613     5734     +121     
==========================================
- Hits         4128     4119       -9     
- Misses       1485     1615     +130     

Files with missing lines	Coverage Δ
pertpy/tools/_distances/_distances.py	85.87% <66.91%> (-4.70%)	⬇️

... and 7 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[selmanozleyen]

Ok I found out why the test fail. I checked myself and when I directly run this cell on the main branch I get Distance.precompute_distances() got an unexpected keyword argument 'verbose'

import matplotlib.pyplot as plt
import numpy as np
import pertpy as pt
import scanpy as sc
from seaborn import clustermap
adata = pt.dt.distance_example()
obs_key = "perturbation"  # defines groups to test

Turns out the notebook was giving this invalid kwarg but it didn't fail because the older implementation would precompute the distances and it wouldn't hit.

if f"{self.obsm_key}_{self.cell_wise_metric}_predistances" not in adata.obsp:
     self.precompute_distances(adata, n_jobs=n_jobs, **kwargs)

while it hits in newer implementation because it doesn't precompute the whole distance matrices in this cell anymore

distance = pt.tl.Distance(metric="euclidean", obsm_key="X_pca")
df = distance.pairwise(adata, groupby=obs_key)

So https://github.com/scverse/pertpy-tutorials/blob/main/distances.ipynb would need updating. That's why I am against kwargs and even if they are used all of the items in there should be checked if they are being passed or not.

[Zethson]

That's why I am against kwargs and even if they are used all of the items in there should be checked if they are being passed or not.

Yes, I agree with you. We can happily change that.

So https://github.com/scverse/pertpy-tutorials/blob/main/distances.ipynb would need updating.

Are you willing to make this change or do you want me to do it? Ideally, we'd include the update commit in this PR.

Thank you!