Skip to content

Latest commit

 

History

History
106 lines (80 loc) · 3.28 KB

CHANGELOG.md

File metadata and controls

106 lines (80 loc) · 3.28 KB
Raw

Change Log

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

Unreleased

[0.3.1] - 2020-03-15

Changed

  • Added upper bound <0.24.0 for scikit-learn version (#59)

[0.3.0] - 2020-10-28

Changed

  • Fix dependencies to make stacking compatible with scikit-learn 0.23+ (#54)
  • Removed support for Python <3.6. (#55)

[0.2.1] - 2020-01-15

Changed

  • Make stacking compatible with scikit-learn v0.22.1. (#52)

[0.2.0] - 2019-12-11

Added

  • Turn on Python 3.7 and 3.8 for Travis CI builds. (#50)

Changed

  • Removed the upper version bound for sklearn. (#50)
  • Update tests and requirements.txt to allow sklearn 0.20 and above. (#47)
  • Instead of boolean flag for dummy_na, have None/False (no dummying), 'expanded' (matches previous True behavior), and 'all' (dummy NAs in all columns where they appear, not just ones we're categorically expanding). (#44)

[0.1.10] - 2019-01-16

Added

  • Raise a RuntimeError if there are more than 5000 levels in a column (#42)
  • Emit a warning if the column levels during transform don't overlap at all with levels from fitting (#41)

[0.1.9] - 2018-05-17

Fixed

  • In DataFrameETL, don't check for levels to expand in columns which are slated to be dropped. This will avoid raising a warning for too many levels in a column if the user has intentionally excluded that column (#39).

[0.1.8] - 2018-04-19

Fixed

  • Fixed DataFrameETL transformations of DataFrames with non-trivial index when preserving DataFrame output type (#32, #33)
  • Add pandas version restrictions by Python version (#37)
  • Fix code which was incompatible with older pandas version (#37)

[0.1.7] - 2018-03-27

Added

  • Added debug log emits for the DataFrameETL transformer (#24, #27)
  • Added debug log emits for the HyperbandSearchCV estimator (#28, #29)
  • Emit a warning if the user attempts to expand a column with too many categories (#25, #26)

[0.1.6] - 2018-1-12

Fixed

  • Now caching CV indices. When CV generators are passed with shuffle=True and no random_state is set, they produce different CV folds on each call to split (#22).
  • Updated scipy dependency in requirements.txt file to scipy>=0.14,<2.0
  • DataFrameETL now correctly handles all Categorial-type columns in input DataFrames. The fix also improves execution time of transform calls by 2-3x (#20).

[0.1.5] - 2017-10-27

Added

  • Added check_null_cols argument to check for null columns (#13)

[0.1.4] - 2017-10-11

Fixed

  • Fixed bug with fit_params handling in stacking (#12)

[0.1.3] - 2017-10-5

Fixed

  • Resolved issues with one and two-level edge cases for categorical expansion (#10)

[0.1.2] - 2017-10-3

Fixed

  • Included y=None in the fit method definition of DataFrameETL (#7)

Changed

  • Improved parallel performance for hyperband (#8)

[0.1.1] - 2017-09-13

Fixed

  • Fixed version requirements for scikit-learn to properly import MaskedArray (#4).
  • In the stacking estimators, get_params no longer throws index error when estimator_list is an empty list (#6).

[0.1.0] - 2017-09-12

Added

  • initial commit