fix: avoid cumulative size reordering across multi-color subplots

[Marius1311]

• Closes Per-point size array is cumulatively reordered across multi-color subplots #4024
• Tests included or not required because: the fix is a one-line variable rename (loop-local _size instead of mutating size); existing tests pass and the bug is visual (per-point size arrays scrambled across subplots).
• Release notes not necessary because: this is a minimal internal bugfix with no API change.

---

When sc.pl.embedding (or sc.pl.umap, etc.) is called with multiple color keys and a per-point size array, the size variable is cumulatively reordered across loop iterations:

# line 299 in scatterplots.py
if isinstance(size, np.ndarray):
    size = np.array(size)[order]  # mutates loop variable!

Each subplot computes its own order (for z-ordering by color value), but applies it to the already-reordered size from the previous iteration. This means:

• Subplot 1: correct sizes
• Subplot 2: sizes scrambled by order1 × order2
• Subplot 3: sizes scrambled by order1 × order2 × order3
• etc.

Fix: use a loop-local _size variable so each iteration reorders from the original size array. The same pattern is applied to the scatter call and add_outline code that consume the size.

Minimal reproduction:

import scanpy as sc
import numpy as np

adata = sc.datasets.pbmc3k_processed()
sizes = np.random.default_rng(0).uniform(10, 200, size=adata.n_obs)
sc.pl.umap(adata, color=["louvain", "n_genes", "n_counts"], size=sizes)
# Dot sizes visibly differ across panels despite being the same array

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.49%. Comparing base (1fbe008) to head (83b4db4).
⚠️ Report is 19 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4023      +/-   ##
==========================================
- Coverage   78.50%   78.49%   -0.01%     
==========================================
  Files         117      117              
  Lines       12740    12739       -1     
==========================================
- Hits        10001    10000       -1     
  Misses       2739     2739              

Flag	Coverage Δ
hatch-test.low-vers	77.79% <100.00%> (-0.01%)	⬇️
hatch-test.pre	77.45% <100.00%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/scanpy/plotting/_tools/scatterplots.py	83.48% <100.00%> (-0.04%)	⬇️

[flying-sheep]

thank you!