GitLab

apparently SWIG typemaps maps Python's list but not set with std::set.
Instead, Python's `set` is mapped with `std::unordered_set`.

The original code of `swigInputBuilder.py` is kindly in a mess. Too much
lines in a single method `writeMethods` which causes debug becomes
difficult. Also there are a lot of warnings about
[PEP8](https://www.python.org/dev/peps/pep-0008/).

A refactoring in the future may be a good idea.

Now that I have my head on straight, that is.

right).  Add a test for AmoebaMultipoleForce.

cubic/quartic/pentic/etc. terms are NOT working for some reason.

expansion coefficients.

docstring says it should be (kJ/nm^2) -- the swig input generator had the
inverse.

Adds a test case for the AmoebaGeneralizedKirkwoodForce API.

angles are in degrees (unlike *every* other angle force out there), and
stripUnits automatically reduces all angles to radians if they come in with
units.

The approach here is to *slightly* change the API, so that
AmoebaAngleForce.addAngle interprets input angles <2*pi as radians, and >2*pi as
degrees. This is heuristic, but should work in every case out in the wild so
far.

I've also updated the documentation to reflect this behavior, and fixed the
units attached to the return value of AmoebaAngleForce.getAngleParameters() to
return degrees instead of radians.

AmoebaBondForce.

Renamed methods for dealing with TabulatedFunctions to avoid problems in wrapper APIs.  Also fixed a few bugs.

Skip all /internal directories when building Python wrappers.  This allows the configuration file to be simplified.