Repository navigation
feat: support composite native pipelines - #16
Merged
Merged
Conversation
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configuration
You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Pipelines ending in XGBClassifier/XGBRegressor or CatBoostClassifier/CatBoostRegressor now export as composite bundles and load as real sklearn Pipelines. The transformer prefix uses the owned skops boundary; the final estimator uses its unchanged native adapter. Both components, fitted widths/names, claims and transformed-coordinate structure are verified before publication and on load.
The strict format-1 recipe declares step names and outer options. Model member pairs require that recipe; the archive ceiling rises to six so schema and explicit samples fit together. Existing standalone wire bytes and all-sklearn pipelines remain unchanged. Unsupported native placements, outer state, metadata routing and unverifiable transformer widths fail on export. No dependency or native type is added to skops trust.
Validation on Python 3.14 after merging current develop (174385a):
Normative rules and rejection messages are in docs/format-pipeline.md. No source, test or documentation file exceeds 300 lines.