fix: replace .max()[0] with .max().iloc[0] for pandas v2.x compatibility

[takuan517]

Summary

Fix a KeyError: 0 crash that occurs across all ML modules when running Planter with pandas >= 2.0.

Root Cause

packages.txt specifies pandas==1.1.3, where .max()[0] happens to work. However, pandas v2.0 removed support for positional integer indexing on Series, so the following pattern raises KeyError: 0 on modern environments (e.g., Ubuntu 24.04 with Python 3.12):

t_t = [test_X[[i]].max()[0], train_X[[i]].max()[0]]

.max() returns a Series indexed by column name, not by position.
Using [0] attempts to look up a column named 0, which does not exist.

Fix

Replace .max()[0] with .max().iloc[0], which uses explicit positional indexing and is compatible with both pandas v1. x and v2. x.

Affected Files

51 table_generator.py files across all ML modules: RF, DT, XGB, SVM, KNN, KM, NN, Bayes, IF, PCA, Autoencoder

Tested

• Python 3.12, pandas 2.x, Ubuntu 24.04 ARM
• RF/Type_EB on Iris dataset: accuracy 0.9556 ✓ (Matrix 1–3 all passing)

[aqn96]

Can confirm this fix works. I independently hit the same KeyError: 0 on DT/Type_4 while running Planter on Ubuntu 24.04 ARM with Python 3.12 and pandas 3.0.1. .max().iloc[0] resolves the crash and results are correct.

Heads up — there are also 6 instances of .min()[0] that have the same breakage on pandas v2.x+:

grep -rn "\.min()\[0\]" src/ --include="*.py"
src/models/KNN/Type_EB/table_generator.py:209
src/models/KNN/Type_1/table_generator.py:208
src/models/KM/Type_LB/table_generator.py:63
src/models/KM/Type_EB/table_generator.py:213
src/models/KM/Type_clustreams/table_generator.py:213
src/models/KM/Type_1/table_generator.py:63

Might be worth including .min()[0] → .min().iloc[0] in this PR as well.

[takuan517]

@aqn96
Thanks for confirming! Good catch on .min()[0] — I've updated the PR to include that fix across the 6 affected files as well.a6c6e34