Move particle generators to their own folder

[TotoGaz]

Remove the dependency (and the transitive dependencies) to SpatialPartition in the mesh/generators "package".

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 71.49758% with 59 lines in your changes are missing coverage. Please review.

Project coverage is 53.23%. Comparing base (9289b76) to head (2564a74).

Files	Patch %	Lines
...Components/mesh/generators/PartitionDescriptor.cpp	78.26%	20 Missing :warning:
...h/particleGenerators/ParticleMeshGeneratorBase.cpp	15.78%	16 Missing :warning:
src/coreComponents/mesh/MeshManager.cpp	59.37%	13 Missing :warning:
...onents/mesh/mpiCommunications/SpatialPartition.cpp	20.00%	4 Missing :warning:
...ents/mesh/generators/InternalWellboreGenerator.cpp	0.00%	3 Missing :warning:
...Components/mesh/generators/PartitionDescriptor.hpp	77.77%	2 Missing :warning:
.../mesh/particleGenerators/ParticleMeshGenerator.cpp	0.00%	1 Missing :warning:

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #3068      +/-   ##
===========================================
+ Coverage    53.19%   53.23%   +0.03%     
===========================================
  Files          989      992       +3     
  Lines        83646    83768     +122     
===========================================
+ Hits         44498    44592      +94     
- Misses       39148    39176      +28     

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[TotoGaz]

Hello @cmcrook5 @homel1

I've tried to split the Particle mesh specials from the standard ones. To do so I've move all the Particle* classes to the particleGenerators folder (plus some other tweaks). The architecture is kind of clearer also: what's dedicated to particles is clearly located.

Still remains this question of SpatialPartition which embeds some specific elements for particles. I've tried to separate by adding a PartitionDescriptor (which also duplicates a little bit of code, which can be acceptable as a temporary solution) It's working for serial mpm simulations, but for parallel, in some cases yes, in other cases no. It's more random. My guess is that in the MeshManager::generateMeshes we transfer the result from the new PartitionDescriptor to the SpatialPartition, but something's wrong in there. Other than that, the failing tests (on the CI) are

FAILED:  3 python3(impermeableFault_smoke_01_2_restartcheck), python3(permeableFault_smoke_01_2_restartcheck), python3(mpm_singleParticle_01_2_restartcheck)
TIMEOUT:  3 geosx(pennyShapedToughnessDominated_smoke_01_1_geos), geosx(pennyShapedViscosityDominated_smoke_01_1_geos), geosx(pknViscosityDominated_smoke_01_1_geos)

which tend to indicate that it's not so bad since currently

FAILED:  2 python3(impermeableFault_smoke_01_2_restartcheck), python3(permeableFault_smoke_01_2_restartcheck)
TIMEOUT:  3 geosx(pennyShapedToughnessDominated_smoke_01_1_geos), geosx(pennyShapedViscosityDominated_smoke_01_1_geos), geosx(pknViscosityDominated_smoke_01_1_geos)

are known to fail.

Is there any chance that you can handle this refactoring? Maybe you can find what needs to be updated. Or if necessary handle some inheritance to deal with your partitioning? We can discuss it when you're ready.

[cmcrook5]

@TotoGaz I pulled your branch and I'm trying to take a look at why the single particle test is failing. However, I'm having some trouble building your branch. I get the follolwing error:

In file included from /usr/WS1/crook5/GEOS_PR/src/coreComponents/linearAlgebra/interfaces/hypre/HypreMGR.cpp:34:
/usr/WS1/crook5/GEOS_PR/src/coreComponents/linearAlgebra/interfaces/hypre/mgrStrategies/SinglePhasePoromechanicsEmbeddedFractures.hpp:117:27: error: use of undeclared
      identifier 'HYPRE_MGRSetFSolverAtLevel'
    GEOS_LAI_CHECK_ERROR( HYPRE_MGRSetFSolverAtLevel( 1, precond.ptr, mgrData.mechSolver.ptr ) );
                          ^
In file included from /usr/WS1/crook5/GEOS_PR/src/coreComponents/linearAlgebra/interfaces/hypre/HypreMGR.cpp:35:
/usr/WS1/crook5/GEOS_PR/src/coreComponents/linearAlgebra/interfaces/hypre/mgrStrategies/SinglePhasePoromechanicsConformingFractures.hpp:124:27: error: use of undeclared
      identifier 'HYPRE_MGRSetFSolverAtLevel'
    GEOS_LAI_CHECK_ERROR( HYPRE_MGRSetFSolverAtLevel( 1, precond.ptr, mgrData.mechSolver.ptr ) );
                          ^

[TotoGaz]

Hi @cmcrook5 thanks you for your actions! Are you using the latest version of the TPLs?

[TotoGaz]

Hello @cmcrook5 did you manage to compile the branch?

[cmcrook5]

@TotoGaz yes, I was able to after updating the TPLs, thanks.