fix(preprocessor): reject an unrecognized scaling value

[ChrisW09]

Fixes #35.

Problem

resolve_method returns its input lowercased when it recognizes nothing, so an unrecognized
scaling value fell through the membership test and produced a pipeline with no scaler:

'minmax'        -> ['imputer', 'minmax', 'ple']
'zscore'        -> ['imputer', 'scaler', 'ple']
'not_a_scaler'  -> ['imputer', 'ple']          <- silently unscaled
'minmaxx'       -> ['imputer', 'ple']          <- silently unscaled

Preprocessor(scaling="minmaxx").fit(df, y) completed without complaint. Meanwhile the same
typo in numerical_method raises, a few lines further down in the same function.

Fix

Validate after alias resolution, keeping None and "none" as the documented ways to turn
scaling off:

InvalidParamError: get_numerical_transformer_steps.scaling = 'minmaxx' is invalid.
Constraint: must name a scaler, or be None / 'none' to disable scaling
Valid values: ['minmax', 'none', 'standardization']

Alias resolution is untouched, so "MinMax", "min-max", "zscore" and
"standardization" all keep working.

Why it matters more than a typical typo guard

The failure is invisible. The fit succeeds, the output widths are identical, and only the
numeric scale of the features differs — so the mistake surfaces, if at all, as a quietly
worse model rather than an error.

Tests

Nine added to tests/test_method_aliases.py: three unknown values that must raise, five
known spellings (including aliases) that must still insert their step, three disabling
spellings that must remain no-ops, and one end-to-end check through the Preprocessor.

Full suite: 492 passed, 9 xfailed. ruff check clean on changed files; pyright unchanged at
its pre-existing 71 errors.

🤖 Generated with Claude Code