Enhancement/md pair gaussian value check

[josephburkhart]

Description

I have modified the TypeParameterDict for md.pair.Gaussian, md.pair.ExpandedGaussian, and md.pair.LJGaussian so that sigma must be a positive real number. Correspondingly, I have updated docstrings and pytests to reflect the new requirement.

Motivation and context

This change ensures that users cannot accidentally or mistakenly set sigma equal to 0 for any of the Gaussian-type pair potentials, since doing so would cause divide-by-zero errors during simulation execution.

Resolves #1806

How has this been tested?

I have added new invalid parameter dictionaries to the constructor function _invalid_params() in hoomd/md/pytest/test_potential.py.

I have built the documentation using Sphinx and confirmed that the changes I made to the docstrings display properly.

Change log

Added a positive-number check to parameters for Gaussian-type pair potentials.

Checklist:

• I have reviewed the Contributor Guidelines.
• I agree with the terms of the HOOMD-blue Contributor Agreement.
• My name is on the list of contributors (sphinx-doc/credits.rst) in the pull request source branch.

[josephburkhart]

@joaander do you have any suggestions on how I can get the pre-commit.ci check to pass? As far as I can tell, I have ensured that my additions a) have no trailing whitespace and b) satisfy the yapf settings in setup.cfg. Am I missing something?

[josephburkhart]

pre-commit.ci autofix

[tommy-waltmann]

Thanks for this, just a small fix in where the positive value is documented.

[joaander]

Thanks!

[tommy-waltmann]

Looks good, thanks for working on this!