fix bug in size_norm extrapolation beyond data range

[lilisako]

Original Issue

• Bug in size_norm extrapolation beyond data range #2539

Behavior

import numpy as np
import pandas as pd
import seaborn as sns
from matplotlib import pyplot as plt

np.random.seed(12345)
df = pd.DataFrame({'x': np.random.rand(60),
                   'y': np.random.rand(60),
                   'sz': np.append(np.random.randint(0, 2500, 30), np.random.randint(2500, 6000, 30))})

fig, axs = plt.subplots(ncols=2, figsize=(10, 4), sharex=True, sharey=True)
sns.scatterplot(data=df[:30], x='x', y='y', size='sz', sizes=(0.01, 200), size_norm=(0, 6000), ax=axs[0])
sns.scatterplot(data=df, x='x', y='y', size='sz', sizes=(0.01, 200), size_norm=(0, 6000), ax=axs[1])

• Before
  

• After
  

[codecov]

Codecov Report

Merging #2579 (3dab64c) into master (4f79753) will increase coverage by 0.19%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2579      +/-   ##
==========================================
+ Coverage   97.27%   97.46%   +0.19%     
==========================================
  Files          17       17              
  Lines        6340     6350      +10     
==========================================
+ Hits         6167     6189      +22     
+ Misses        173      161      -12     

Impacted Files	Coverage Δ
seaborn/rcmod.py	100.00% <ø> (ø)
seaborn/_core.py	98.61% <100.00%> (-0.01%)	⬇️
seaborn/categorical.py	98.53% <100.00%> (+0.14%)	⬆️
seaborn/matrix.py	95.87% <100.00%> (+<0.01%)	⬆️
seaborn/regression.py	97.88% <100.00%> (+0.71%)	⬆️
seaborn/distributions.py	96.37% <0.00%> (+0.32%)	⬆️
seaborn/_docstrings.py	100.00% <0.00%> (+12.50%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 536fa2d...3dab64c. Read the comment docs.

[mwaskom]

Thanks for tracking down the root cause here! I pulled your commits into #2580 and built on them a little bit to add a test for the original problem and to reorganize things a bit so that the _default_size_range attribute was not accessed in a new place (I would like to clean up how that works because it is a bit of a mess at the moment).

PS its' generally cleaner to create a new branch before making edits rather than adding commits onto the master branch of your fork.

Thanks again!

[lilisako]

Thanks for the review & merging my commits! I'll make sure to create a new branch next time.