Automunge

Latest version: v8.33

Safety actively analyzes 715032 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 96 of 99

2.29

- added a test for nan for binarymissingplug in process binary class
- updated logic tests in evalcategory to address potential assignment scenarios to bnry or text class
- fixed error in evalcategory (an elif was used instead of an if)
- fixed another error in evalcateogry, changed mc2 from mc2[0][0] to mc2[1][0]
- small correction in PCA methods archiving parameters
- printouts for feature importance and pCA dimensionality reduction
- final printouts

2.28

- made powertransform evaluation optional to be more consistent with documentaiton
- changed order of operations for label column to before infill to be consistent with documentaiton
- fixed infill loop to be consistent with documentation
- updated zeroinfill to work on multicolumn sets
- added new printouts to signal completion

2.27

additional printstatus printout before category evaluation to aid troubleshooting in operation

2.26

found a typo in evalcategory

2.25

- noted that evaluation of numerical data now includes method to apply transformations based on inferred distribution properties
- updated the family tree primitives based on removal of grandparents, great-grandparents to address discrepancy in documentation (equivalent functionality was possible with fewer primitives)
- added description for MAD3 processing category added to library with 2.25

2.24

- added check for assigncat and assigninfill to return error message if redundant column assignments
- checked Null determination in evalcategory from >= 85% NaN to >80% NaN.
- unique value evaluation in evalcategory now only performed once and saved
- added a note at header for livening /. intellectual property
- cleanup inn postprocess_bxcx
- fixed bug with shuffle train recently introduced as df_labels was referenced before assignment
- silenced evaluation for train category and test category comparison for cases where no df_test was passed
- changed the use of numbercategoryheuristic now it is only used for determination of ordinal processing
- fixed bug with process_mnm3
- changed transformdict definition of MAD2
- fixed bxc4 trasnformdict definition
- fixed scenario where test column only has two categories (bury) and train has more for either text or ordl
(necessary when deferring to evalcategory for category)
- bug fix for postprocess_date applicable to dat2
- added new argument to insertinfill function to handle different use in adjinfill when only inserting infill
to one column at a time for a multi column set
- performed audit of processing function root categories

Page 96 of 99

Links

Releases

Has known vulnerabilities

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.