Skip to content

direpack_Release_Notes.md

Latest commit

 

History

History
123 lines (91 loc) · 4.76 KB

direpack_Release_Notes.md

File metadata and controls

123 lines (91 loc) · 4.76 KB

sprm Release notes (versions 0.0 through 0.7)

Version 0.2.1

  • sprm now takes both numeric (n,1) np matrices and (n,) np.arrays as input

Version 0.2.0

Changes compared to version 0.1:

  • All functionalities can now be loaded in modular way, e.g. to use plotting functions, now source the plot function separately:

      from sprm import sprm_plot

  • The package now includes a robust M regression estimator (rm.py), which is a multiple regression only variant of sprm. It is based on the same iterative re-weighting scheme, buit does not perform dimension reduction, nor variable selection.

  • The robust preprocessing routine (robcent.py) has been re-written so as to be more consistent with sklearn.

Version 0.3

All three estimators provided as separate classes in module:

    from sprm import sprm 
    from sprm import snipls
    from sprm import rm

Also, sprm now includes a check for zero scales. It will remove zero scale variables from the input data, and only use columns corresponding to nonzero predictor scales in new data. This check has not yet been built in for snipls or rm separately.

Plus some minor changes to make it consistent with the latest numpy and matplotlib versions.

Version 0.4

The preprocesing routine robcent has been refactored. Functionality has been added to centre the data nonparametrically by the L1 median. The ancillary functions for robcent have been moved into _preproc_utilities.py.

Furthermore, sprm, snipls and rm have all three been modified such that they accept matrix, array or data frame input for both X and y. Also, the option to provide column names has been extended to automatic extraction from data frame input, or direct input as list, array or pandas Index.

The license has been changed from GPL3 to MIT.

0.4.2. 'kstepLTS' location estimator included.

Version 0.5

Pre-processing functions further refactored so as to be compatible with sklearn pipelines. Class now named VersatileScaler; the old robcent name still works, but will be sunset.

Version 0.6

Preprocessing files moved into separate folder. More preprocessing options. Examples moved into Jupyter notebook in separate examples section.

direpack release notes (since version 0.8)

Version 0.8

ppdire merges in

Version 0.9

  • preprocessing widely extended
  • plot functions adapted
  • documentation improved

Version 1.0

  • sudire joins in
  • plot functions adapted
  • documentation provided for dicomo
  • 1.0.2: link to direpack publication added
  • 1.0.3: fixed rare division by zero in l1median
  • 1.0.4: unit tests included
  • 1.0.5: sudire notebook adapted
  • 1.0.9: function to calculate the martingale difference divergence matrix (MDDM) added in _dicomo_utils.py
  • 1.0.11: documentation updated to accommodate for go-live of readthedocs page
  • 1.0.13: fixed bug in option to use Ball in sudire. Adjusted readthedocs.

Version 1.2

Updated to Python 3.10-3.12 standards and NumPy 2.0 compatibility.

CI/CD Updates:

  • Updated GitHub Actions to latest versions (actions/checkout@v4, setup-miniconda@v3)
  • Updated Python test matrix from 3.8-3.10 to 3.9-3.12
  • Fixed setup.py to read version without importing package (avoids dependency issues during install)

NumPy 2.0 Compatibility:

  • Replaced deprecated np.matrix with np.asarray throughout codebase and notebooks
  • Fixed np.product → np.prod (removed in NumPy 2.0)
  • Fixed np.asfarray → np.asarray (removed in NumPy 2.0)
  • Fixed float() conversion on non-0D arrays using .item() or proper indexing
  • Fixed matrix multiplication in ppdire.transform() using np.dot() instead of * operator

Python 3.12 Compatibility:

  • Fixed deprecated assertAlmostEquals → assertAlmostEqual in test files

cyipopt/IPOPT Changes:

  • Made cyipopt an optional dependency: standard install with pip install direpack, full install with pip install direpack[ipopt]
  • Fixed cyipopt API changes: cyipopt.problem → cyipopt.Problem, addOption → add_option
  • sudire module only available when cyipopt is installed
  • Added IPOPT_AVAILABLE flag to check availability at runtime

Bug Fixes:

  • Fixed sprm_plot_cv.eta_ncomp_contour() to handle NaN values in cross-validation results
  • Fixed sklearn PLSRegression.coef_ transposition in example notebooks (API changed in sklearn 1.3)
  • Fixed edge cases in _sudire_utils.py for slices with ≤1 sample

Documentation:

  • Updated README with two installation options (standard and full with ipopt)
  • Updated example notebooks (dicomo_example.ipynb, ppdire_example.ipynb) for NumPy 2.0 compatibility
⚡