[Wavepacket-user] Wavepacket (Python) 0.4 released

[Ulf L. <ul...@wp...>]

Hello everyone,

I have just pushed the button for releasing the latest Python
package for Wavepacket 0.4.

You can get the latest release from PyPI, e.g., with
"pip install wavepacket"


The main goals were getting over with most breaking changes so that the
API is stable afterwards, and to start with proper support for
two-dimensional problems.


Breaking changes
----------------

* Moved a lot of functionality around to improve consistency and
  explorability
    - Generators (Gaussian, SphericalHarmonics etc.) moved to a
      subpackage 'special', similar to scipy
    - general-purpose functions (trace, expectation_value, ...) moved
      to main namespace
* swapped arguments for plotting functions to be consistent with
  the return value of solver.propagate()


Major new features
------------------

* Added contour plots for 2D systems
* Calculating the density also allows calculation of reduced density
* Gaussian generator can also be used for pulse shapes
* More copying of user data for robustness and some better error
  messages


New docs
--------

* Convergence issues of equally-spaced grids and absorbing boundaries
  https://wavepacket.readthedocs.io/stable/advanced/plane_wave_grid.html
* Convergence issues with ODE solvers
  https://wavepacket.readthedocs.io/stable/advanced/ode_solvers.html


Bug fixes
---------

* 1D plots calculated abs(energy) instead of real(energy)
  and treated unnormalized wave functions incorrectly.
* Population calculation lacked an abs(), so populations could be
  negative
* SoftRectangularFunction had an incorrect condition that effectively
  suppressed the soft turn-off
* Random wave functions were constructed in the wrong representation and
  had wrong normalization, but could be completely wrong in principle.