Repository navigation
Conversation
skrub changes that matter for stratum: - Perf: building DataOps and SkrubLearners over deep graphs is much faster; skrub's graph traversals now memoize node results (skrub#2296, 0.11) - .skb.apply(): `how=` replaced by `no_wrap: bool` plus new `exclude_cols`; Apply also gains `no_cache` (persistent caching, skrub#2017, 0.11) - Non-final estimators: an Apply whose output feeds another Apply now transform()s/predict()s in every response pass instead of answering the scorer (skrub#2052, 0.9) - TableVectorizer converts duration columns to float seconds via the new DurationToFloat step (skrub#2069, 0.10) - .skb.concat() accepts numpy arrays (skrub#2096, 0.10) - skrub.var(..., becomes_default=True): an env that does not bind the var falls back to its value (skrub#2082, 0.10) - .skb.with_scoring() declares scorers inside the DataOp, kwargs may be DataOps such as sample_weight (skrub#1995, 0.9) - .skb.train_test_split() with a declared cv returns the last split (skrub#2213, 0.10.1) - Regression: ApplyToCols.transform() no longer forwards transform kwargs Stratum follow-ups: - Estimator ops carry exclude_cols/no_wrap instead of how - New feeds_estimator flag gives skrub's non-final-estimator semantics - Fused TableVectorizer adds DurationToFloat (pandas/polars parity) - ConcatOp concatenates numpy arrays at runtime - SelectCols -> ColumnSelectorOp only when the Apply gets every column - Unbound becomes_default vars resolve to their value - with_scoring metrics used by grid_search, kwargs evaluated per test fold - IRNode.detach()/detach_and_prune() replace hand-rolled unlinking; fixes a dangling-subgraph crash in the x*0 / x**0 folds
Codecov Report❌ Patch coverage is
Flags with carried forward coverage won't be shown. Click here to find out more.
🚀 New features to boost your workflow:
|
e-strauss
added a commit
to e-strauss/stratum-fork
that referenced
this pull request
Oct 6, 2026
skrub changes that matter for stratum: - Perf: building DataOps and SkrubLearners over deep graphs is much faster; skrub's graph traversals now memoize node results (skrub#2296, 0.11) - .skb.apply(): `how=` replaced by `no_wrap: bool` plus new `exclude_cols`; Apply also gains `no_cache` (persistent caching, skrub#2017, 0.11) - Non-final estimators: an Apply whose output feeds another Apply now transform()s/predict()s in every response pass instead of answering the scorer (skrub#2052, 0.9) - TableVectorizer converts duration columns to float seconds via the new DurationToFloat step (skrub#2069, 0.10) - .skb.concat() accepts numpy arrays (skrub#2096, 0.10) - skrub.var(..., becomes_default=True): an env that does not bind the var falls back to its value (skrub#2082, 0.10) - .skb.with_scoring() declares scorers inside the DataOp, kwargs may be DataOps such as sample_weight (skrub#1995, 0.9) - .skb.train_test_split() with a declared cv returns the last split (skrub#2213, 0.10.1) - Regression: ApplyToCols.transform() no longer forwards transform kwargs Stratum follow-ups: - Estimator ops carry exclude_cols/no_wrap instead of how - New feeds_estimator flag gives skrub's non-final-estimator semantics - Fused TableVectorizer adds DurationToFloat (pandas/polars parity) - ConcatOp concatenates numpy arrays at runtime - SelectCols -> ColumnSelectorOp only when the Apply gets every column - Unbound becomes_default vars resolve to their value - with_scoring metrics used by grid_search, kwargs evaluated per test fold - IRNode.detach()/detach_and_prune() replace hand-rolled unlinking; fixes a dangling-subgraph crash in the x*0 / x**0 folds Closes deem-data#241.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
skrub changes that matter for stratum:
how=replaced byno_wrap: boolplus newexclude_cols; Apply also gainsno_cache(persistent caching, skrub#2017, 0.11)Stratum follow-ups: