FIX: make HuberRegressor accept boolean X

[casey-brooks]

Summary

• coerce inputs to accepted FLOAT_DTYPES via check_X_y to avoid boolean negation errors in _huber_loss_and_gradient
• add dense and sparse regression tests that compare boolean matrices with their float-cast counterparts
• capture the legacy failure stack trace for reference

Pre-fix failure

Traceback (most recent call last):
  File "<stdin>", line 7, in <module>
  File "sklearn/linear_model/huber.py", line 284, in fit
    parameters, f, dict_ = optimize.fmin_l_bfgs_b(
  File "scipy/optimize/lbfgsb.py", line 197, in fmin_l_bfgs_b
    res = _minimize_lbfgsb(fun, x0, args=args, jac=jac, bounds=bounds,
  File "scipy/optimize/lbfgsb.py", line 306, in _minimize_lbfgsb
    sf = _prepare_scalar_function(fun, x0, jac=jac, args=args, epsilon=eps,
  File "scipy/optimize/optimize.py", line 261, in _prepare_scalar_function
    sf = ScalarFunction(fun, x0, args, grad, hess,
  File "scipy/optimize/_differentiable_functions.py", line 140, in __init__
    self._update_fun()
  File "scipy/optimize/_differentiable_functions.py", line 233, in _update_fun
    self._update_fun_impl()
  File "scipy/optimize/_differentiable_functions.py", line 137, in update_fun
    self.f = fun_wrapped(self.x)
  File "scipy/optimize/optimize.py", line 74, in __call__
    self._compute_if_needed(x, *args)
  File "scipy/optimize/optimize.py", line 68, in _compute_if_needed
    fg = self.fun(x, *args)
  File "sklearn/linear_model/huber.py", line 93, in _huber_loss_and_gradient
    X_non_outliers = -axis0_safe_slice(X, ~outliers_mask, n_non_outliers)
TypeError: The numpy boolean negative, the `-` operator, is not supported, use the `~` operator or the logical_not function instead.

Testing

• pytest -q sklearn/linear_model/tests/test_huber.py --maxfail=1
• git diff -U0 | flake8 --diff

Fixes scikit-learn#32949

[casey-brooks]

Test & Lint Summary

• pytest -q sklearn/linear_model/tests/test_huber.py --maxfail=1 (11 passed)
• git diff -U0 | flake8 --diff (no findings)

[noa-lucent]

Fix ensures HuberRegressor converts boolean inputs to float via check_X_y and adds dense/sparse regression tests. Looks good.