[WIP] Removed assert and allowed bilinear filtering along r with RZ FDTD

[oshapoval]

trafficstars

• To do: we might want to remove the charge buffer from WarpX within a separate PR, which should also fix one of the failing CI tests.

This PR removes an abort message which prevented from using bilinear filtering for charge and currents on the mesh, along r with RZ FDTD. Previously this assert message was added because we had observed strange results in RZ FDTD when filtering was ON ( see #1943 ) that we do not observe anymore.

Now, the default is warpx.use_filter=1 in all cases, including the simulations with RZ FDTD solver. Note, with the RZ PSATD solver, the filtering is done k-space. One can also set the number of passes along each direction warpx.filter_npass_each_dir (3 int), and default is 1 1 1.

Here is the corresponding input script:

inputs_rz.txt

[ax3l]

@oshapoval can you locally take a look why ElectrostaticSphereEB_RZ_MR crashes pls?