Release v0.35.0

[chukarsten]

v0.35.0 Oct. 14, 2021

Enhancements

• Added human-readable pipeline explanations to model understanding Human Readable Pipeline Explanations #2861
• Updated to support Featuretools 1.0.0 and nlp-primitives 2.0.0 Upgrade to support Featuretools 1.0.0 and nlp-primitives 2.0.0 #2848

Fixes

• Fixed bug where long mode for the top level search method was not respected Add max batches stopping criterion logic #2875
• Pinned cmdstan to 0.28.0 in cmdstan-builder to prevent future breaking of support for Prophet Update cmdstan to 0.28.0 #2880
• Added Jarque-Bera to the TargetDistributionDataCheck Add Jarque-Bera #2891

Changes

• Updated pipelines to use a label encoder component instead of doing encoding on the pipeline level Use label encoder component instead of doing encoding on the pipeline level #2821
• Deleted scikit-learn ensembler Delete scikit-learn ensembler #2819
• Refactored pipeline building logic out of AutoMLSearch and into IterativeAlgorithm Move pipeline building into IterativeAlgorithm #2854
• Refactored names for methods in ComponentGraph and PipelineBase Refactor ComponentGraph methods #2902

Documentation Changes

• Updated install.ipynb to reflect flexibility for cmdstan version installation Update cmdstan to 0.28.0 #2880
• Updated the conda section of our contributing guide Update conda section of contributing guide #2899

Testing Changes

• Updated test_all_estimators to account for Prophet being allowed for Python 3.9 Update test_all_estimators to account for Prophet in 3.9 #2892
• Updated linux tests to use cmdstan-builder==0.0.8 Update cmdstan to 0.28.0 #2880

Breaking Changes

• Updated pipelines to use a label encoder component instead of doing encoding on the pipeline level. This means that pipelines will no longer automatically encode non-numerical targets. Please use a label encoder if working with classification problems and non-numeric targets. Use label encoder component instead of doing encoding on the pipeline level #2821
• Deleted scikit-learn ensembler Delete scikit-learn ensembler #2819
• IterativeAlgorithm now requires X, y, problem_type as required arguments as well as sampler_name, allowed_model_families, allowed_component_graphs, max_batches, and verbose as optional arguments Move pipeline building into IterativeAlgorithm #2854
• Changed method names of fit_features and compute_final_component_features to fit_and_transform_all_but_final and transform_all_but_final in ComponentGraph, and compute_estimator_features to transform_all_but_final in pipeline classes Refactor ComponentGraph methods #2902

[codecov]

Codecov Report

Merging #2918 (ac5255c) into main (5c85614) will not change coverage.
The diff coverage is 100.0%.

@@          Coverage Diff          @@
##            main   #2918   +/-   ##
=====================================
  Coverage   99.7%   99.7%           
=====================================
  Files        302     302           
  Lines      28396   28396           
=====================================
  Hits       28303   28303           
  Misses        93      93           

Impacted Files	Coverage Δ
evalml/__init__.py	100.0% <100.0%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5c85614...ac5255c. Read the comment docs.

[freddyaboulton]

Perf tests look good to me except the SeoulBikes error because of ww 0.8.1. Thanks @chukarsten !

I say we ship this release and plan to do a minor release next week when #2909 is merged?

What do people think?

[chukarsten]

Absolutely, @freddyaboulton .