WIP: handling of negative poloidal field

[bendudson]

Checking calculations for negative poloidal magnetic field (Jacobian < 0)
Following checking of the coordinate system in boutproject/BOUT-dev#2165

TODO

• Metric components
  • Test with forward & reverse fields (Bpol, Btor)
• zShift, TwistShift
  • Test with forward & reverse fields (Bpol, Btor)
• Curvature components
  • Test with forward & reverse fields (Bpol, Btor)

[johnomotani]

Possible alternative to this PR that might be simpler to implement: modify hypnotoad/cases/tokamak.py so that if 'reversed current' is detected, the gridding starts from the outer, lower divertor and winds anti-clockwise. That would mean that the y-direction on the grid is always parallel to the magnetic field (not anti-parallel), which is probably less likely to suffer from sign errors than fully supporting 'negative Bp'.

[bendudson]

Hey @johnomotani That sounds like a good option to have, at least as an alternative to a negative-Jacobian coordinate system. Even having that capability would allow us to compare simulation results with and without a reversed (anticlockwise) grid to check that these coordinate system issues are handled correctly.