+ Although the parameter is named maxh in netgen,
+ the maximum element diameter of the mesh violates this
+ restriction. Not seldom by more than 100%...
+ Reproduce, e.g., with unit cube and maxh=1

+ Solution: choose maxh= meshSize / j for j=1,2,... until satisfied

+ Should fix problems with bem++
+ to be tested with sp4

+ slonczewski + oersted field now supported for all tp-schemes
+ oersted field is not recomputed if je = const(in time)

+ Adapted an old implementation to new syntax / commix
+ This is just a basic working version for now

+ default alphaPreconditioner=None --> Use alpha for preconditioner
+ JacobiUpdate is skipped for TPS1(PF)
+ moved stuff between _tpsx.py _intergrator.py tps1.py

+ renamed alphaP --> alphaPreconditioner
+ removed various prints and comments
+ removed gmres restart number 200 in tps1

- i.e. for strayField, oerstedField or magnetostaticField(=both)
- moved ComputeMaxwell to Magnetization class
- added numthreads to the test files
- added some missing import sys

+ for all tangent plane schemes
+ for TPS2 and TPS2AB the analysis requires Je in C^2([0,T])

+ now they all derive from the same base class
+ was necessary to not initialize numthreads or need mpi
+ additionally commmented out all bounding box (TOECDO) stuff

+ moved all imports from top of files to functions/methods using them
+ using os.environ from within python to set environment variables for interpreter
+ use by: I = Integrator(..., numthreads=42)

+ TODO: tps2(ab) not correct Je'(t) ignored