Refactor Projection Argument and Time Metadata; Fix Kajiura Filter and Coordinate Arrays

[Thomas-Ulrich]

• Change argument CRS_reference to projection, because we should not restrict to transverse Mercator projections.
• Change time metadata (seconds) of NetCDF, else I'm not able to visualize it in Paraview.
• Fix the Kajiura filter (else it was plotting only the filtered transient displacement increment and not displacement.
• Fix missing +1 in exchangeGrid/exchangeGridCreation.py
• Fix the coordinate arrays that were shifted to start at 0 in donorModels/seissol2exchangegrid.py (I don't understand why this shifting was done).

Now I can run the Madden benchmark setup with:

python interface_module.py -d seissol ~/trash/output_AscSimple_dip16_withCoh_PfLith_mu4_scenario_A2-surface.xdmf /home/ulrich/clawpack/geoclaw/examples/tsunami/MaddenBenchmark/bath2000.nc --resolution 1000 --filter kajiura --filtering_depth 2000

And here is a comparison of the filter data for left SAMPLER (SeisSol/SAMPLER#12), right interface module, same resolution:

(note that the Kaijiura is applied everywhere on SAMPLER, also on land, which explains the differences to the right).

[mbaensch]

Check whether the x_tmp/y_tmp removal in seissol2exchangegrid moves the fault location / location of highest deformation if the spatial resolution is changed. Everything else is a good addition.
The transformation might need the values [0,0] for the lower left corner and x_tmp/y_tmp make sure of that.

[valemagni]

Tried with the Samos example and the fault seems to be in the right place.
This is the last time step of the deformation coming out from the Interface Module:

In this case, the CRS reference is needed. This is the command I used to run the interface module:
python interface_module.py --donor seissol --projection "+proj=tmerc +datum=WGS84 +k=0.9996 +lon_0=26.7953 +lat_0=37.8973" simulations_seissol/dyn_coh0.25_2.5_B1.2_C0.3_R0.65_extracted-surface.xdmf bathySamos.grd --casename bathySamos.
@Thomas-Ulrich I would suggest to add an explenation in the README on how to use the CRS flag now, since the input is not coordinates any longer and give an example. Then I think this can be merged.