Version 0.17.1 (12th of May 2021)

Note

Intermediate release to solve an Issue with the LmFitSolver that is experienced by many users. See Issue 295.

Version 0.17 (2nd of April 2021)

This version primarily fixes a number of Bugs and introduced some performance improvements. This version should be backward-compatible with the previous release.

Note

The Pastas.Project class has been removed from Pastas in this release. Pastastore is now recommended to deal with multiple time series models.

New Features / Enhancements

  • ml.set_parameter now also takes an optimal argument.

  • ml.get_stressmodel_settings has been added. This method can be used to obtains the TimeSeries settings in a stressmodel.

  • Predefined TimeSeries settings can now be accessed through ps.rcParams

  • The Parameters of the response function ps.Hantush have been rescaled such that parameter ‘A’ is the gain.

  • The metrics AIC and BIC now use the noise when available, and residuals otherwise.

  • The use of f-strings is now the preferred method for strings. For the logger messages s-strings are now consistently applied to further speed-up the code.

  • The method ml.plots.cum_frequency is added to the plotting library.

Deprecations

No methods were deprecated in this version.

Backwards incompatible API changes

  • pastas.Project class is removed from pastas. Use Pastastore instead.

  • ps.FactorModel stressmodel is removed. Use ps.StressModel with ps.One response function instead.

New Example Notebooks

  • A new example notebook on model calibration has been added.

  • A new example notebook on recharge estimation has been added.

  • A new example notebook on changing the time step has been added.