GitLab

errors.

This copies the fix for the AmoebaTorsionTorsionForce over to the other two
forces that calculate angles via the inverse cosine -- specifically the in-plane
angle force and the stretch-bend force.  The TorsionTorsion fix did not
completely solve the occasional simulation blowup, although it did make it a
*lot* less frequent (it went from happening within 5 minutes to happening ~once
per day).

The forces were not correct for different uind/uinp as a result.

Non-PME OPT calculations now use a much more efficient Cartesian field algorithm for the dipole response force terms.

These are worth fixing because they're in header files that are always included
(rather than just hitting them at OpenMM compile-time). The SWIG interface now
builds without warnings on my machine.

Swigged python docstrings now include documented return values and type
information or their arguments. They are generated in numpydoc format.
Furthermore, all of the Python app layer docstrings have been changed
to numpydoc format. The filterPythonFiles.py script which helps to
generate the Doxygen Python API docs has been updated to reflect these
changes.

New advanced options OPENMM_BUILD_CPU_TESTS, OPENMM_BUILD_SERIALIZATION_TESTS,
OPENMM_BUILD_REFERENCE_TESTS were added to the cmake configuration and disabled
in the OpenCL test matrix element to increase speed. Redundant or unnecessary
apt packages removed from the travis installation.

- Remove API change for AmoebaAngleForce so it goes back to only taking degrees.
  Handle conversion via a pythonprepend (since stripUnits is done in C++ now).
- Do the same thing for TorsionTorsionForce, since angles are in degrees again
- Add AmoebaVdwForce and AmoebaWcaForce tests

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.

Specified units for getOPTCoefficients() routine and switched µ for mu, to avoid unicode writing errors.

The QI direct space energy uses the energy functional

    U = 1/4 (E_p µ_d + E_d µ_p)

while the code it replaced uses the equivalent form

    U = 1/2 E_p µ_d

Therefore the reciprocal and self terms have been adapted to use the more symmetric form.  The bad code yielded incorrect energies when PME is used AND µ_d and µ_p are different.  The forces and torques were not affected.