fix: remove dead branch in WealthPercentileTransformer.fit (#168)

[Larslllllll]

Summary

The hasattr(X, "columns") branch in fit() never executed because validate_data() returns a NumPy array, which never has a .columns attribute. validate_data() itself sets feature_names_in_ when given a DataFrame, so the column names are preserved regardless.

The elif branch (array-input path) was doing all the work. This PR deletes the dead branch and promotes the elif to a plain if with an explanatory comment.

Changes

• Deleted the dead if hasattr(X, "columns"): branch
• Promoted the elif to a plain if with comment explaining why the check is on self not X
• Added test test_wealth_percentile_feature_names_match_for_frame_and_array_input to pin expected feature names for both DataFrame and array inputs

Verification

# Verify dead branch is gone
test "$(grep -c 'hasattr(X, "columns")' philanthropy/preprocessing/_wealth_percentile.py)" = "0"

# Verify behavior is preserved
python -c "
import numpy as np, pandas as pd
from philanthropy.preprocessing import WealthPercentileTransformer
df = pd.DataFrame({'net_worth': [1.0, 2.0, 3.0], 'other': [1.0, 2.0, 3.0]})
a = WealthPercentileTransformer().fit(df)
b = WealthPercentileTransformer().fit(df.to_numpy())
assert list(a.get_feature_names_out()) == ['net_worth', 'other', 'net_worth_pct_rank']
assert list(b.get_feature_names_out()) == ['x0', 'x1']
"

Fixes #168

[shivamlalakiya]

Thank you, @Larslllllll. Closing this one as already-landed rather than rejected, and your credit is safe on both counts.

Where each piece went:

• The edge-case tests here (all-missing and partially-missing wealth columns) were merged as test: add edge-case tests for WealthPercentileTransformer missing columns (#169) #174 and are on main now, along with your CONTRIBUTORS.md entry.
• The dead hasattr(X, "columns") branch removal was merged as part of fix: reject missing wealth columns #179 (6dfafa5), which closed cleanup: WealthPercentileTransformer.fit has a branch that can never run #168 alongside WealthPercentileTransformer silently ignores a wealth_cols list that matches no column #156.

So everything in this PR is on main, which is also why it now shows as conflicting: it is trying to re-add lines that are already there.

The duplication with #175 the same evening was our process failure, not yours. A "claim the issue before you open a PR" rule is going into CONTRIBUTING.md so contributors stop colliding like this.

Closing as already merged via #174 and #179.