Skip to content

Fable - #17

Open
jsture wants to merge 9 commits into
mainfrom
fable
Open

Fable#17
jsture wants to merge 9 commits into
mainfrom
fable

Conversation

@jsture

@jsture jsture commented Jul 13, 2026

Copy link
Copy Markdown
Contributor

No description provided.

@jsture jsture self-assigned this Jul 13, 2026
jsture added 6 commits July 13, 2026 10:28
Align O2PLS.fit with the sklearn estimator contract (sklearn's own
cross_decomposition deprecated Y in favour of y). Positional calls
unaffected; keyword calls must use fit(X, y=...). Y-block helpers
(predict_x, transform_y, ...) keep their uppercase Y block argument.

Remove O2PLS.score: it duplicated RegressorMixin.score exactly; the
method remains available via the mixin.

Also complete the O2PLS Attributes docstring (b_t_/b_u_, filtered and
residual training blocks with a pickle-size note, fitted component
counts, n_features_in_/feature_names_in_) and note that the string
scale parameter differs from PLSRegression's boolean scale.
Document previously undocumented fitted attributes per the sklearn
docstring standard: OPLS per-component diagnostics (r2x_components_,
r2x_ortho_components_, r2y_components_), training Q residuals and
residual sums of squares, and n_features_in_/feature_names_in_ on both
estimators plus n_orthogonal_ on OPLSDA. Note on OPLS that the string
scale parameter differs from PLSRegression's boolean scale.
Replace np.testing.assert_allclose with
sklearn.utils._testing.assert_allclose, which the sklearn developer
guide recommends for continuous-value comparisons (it infers rtol from
the array dtypes). Exact-equality assert_array_equal calls unchanged.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant