All pull requests

Pull requests

• fix(transformers): record and check DataFrame feature names

  #135·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(serialization): write specs as strict JSON without NaN tokens

  #134·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(serialization): encode object arrays and builtin dtypes in specs

  #132·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(compose): pass target_aware to optionally supervised representations

  #131·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): keep a block for columns with no observed value at fit

  #130·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): accept polars DataFrames as input

  #129·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(validation): reject an unknown task instead of using classification

  #128·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(search): make the search a classifier when it wraps a classifier

  #127·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(search): pass groups to the splitter and fit params to the estimator

  #126·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(cross-fitting): return transform's output type from fit_transform

  #125·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(cross-fitting): keep missing indicators of a wrapped Preprocessor

  #124·ChrisW09 opened Oct 16, 2026Collaborator·

  ·

• fix(splines): ignore y when unsupervised and reject a multi-output target

  #123·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(splines): expand a missing value to a NaN row in the B/M/I and thin-plate bases

  #122·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(splines): keep family policy defaults and extrapolate out-of-range rows

  #121·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(feature-maps): drop the unsupervised relu center at the data maximum

  #120·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(placement): return distinct quantile locations on tied data

  #119·ChrisW09 opened Oct 16, 2026Collaborator·

  ·

• fix(splines): honour explicit knot_locations regardless of output_dim

  #118·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(cross-fitting): refit only target-aware Preprocessor blocks per fold

  #117·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): match array input by position after a named fit

  #116·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(search): default to the classification task for a classifier

  #115·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): keep the internal ColumnTransformer on default output

  #114·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): resolve output_format='auto' once per fit

  #113·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): reject NumPy input of the wrong width at transform

  #112·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(preprocessor): keep ColumnTransformer step names valid for any label

  #111·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1

• fix(placement): rank CART splits by impurity decrease before spacing

  #110·ChrisW09 opened Oct 16, 2026Collaborator·

  ·
  1