coarsening test

[nicolasaunai]

Coarsening test WithRatioFrom2To10TestThat/ALinearFieldCoarsen1DO1.conserveLinearFunction/0 is broken with MPI.

• do we actually need that test if sub-components are tested?
• can we test the coarsening is OK for all quantities and all interp_order and all dimensions in a functional test?

Come back on that once #58 is done

[nicolasaunai]

ALinearFieldCoarsen1DO1 This test uses a 2 levels basic hierarchy where each quantity is intialized on level 1 to a linear function and checks that after applying a coarsening schedule, values on the coarse level are correct. This test is only 1D. It is also failing in MPI probably because it assumes that user refinement boxes are the patches on the refined level...

This test could be re made in python by initializing a hierarchy, advancing a time step and dumping diags. Advancing 1 coarse step would make a sync of the refined level and we could check coarse values where there are fine patches more easily from a python hierarchy. once done in python the cpp test can be deleted.

for each refined level in the hierarchy, for each patch, assume the values of each field (EM and moments) on the patch is the ground truth and check if the values on the underlying next coarser patch are consistent with what the coarsening operation should give.