Add pressure Diag #769
Add pressure Diag #769
Conversation
|
on top of #766 |
WalkthroughWalkthroughThe changes primarily involve the introduction of a new Changes
TipsChat with CodeRabbit Bot (
|
nicolasaunai
force-pushed
the
pressurereb
branch
from
f1fe206 to
4b2f422
Compare
nicolasaunai
force-pushed
the
pressurereb
branch
3 times, most recently
from
0272592 to
2410004
Compare
nicolasaunai
force-pushed
the
pressurereb
branch
from
2410004 to
15730ea
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Review Status
Actionable comments generated: 13
Configuration used: .coderabbit.yaml
Files selected for processing (26)
- src/amr/data/field/coarsening/field_coarsen_operator.hpp (2 hunks)
- src/core/CMakeLists.txt (1 hunks)
- src/core/data/grid/gridlayoutdefs.hpp (1 hunks)
- src/core/data/grid/gridlayoutimplyee.hpp (4 hunks)
- src/core/data/ions/ion_population/ion_population.hpp (8 hunks)
- src/core/data/ions/ions.hpp (9 hunks)
- src/core/data/tensorfield/tensorfield.hpp (1 hunks)
- src/core/data/vecfield/vecfield.hpp (2 hunks)
- src/core/data/vecfield/vecfield_component.hpp (1 hunks)
- src/core/hybrid/hybrid_quantities.hpp (2 hunks)
- src/core/numerics/interpolator/interpolator.hpp (8 hunks)
- src/diagnostic/detail/h5writer.hpp (14 hunks)
- src/diagnostic/detail/types/electromag.hpp (2 hunks)
- src/diagnostic/detail/types/fluid.hpp (10 hunks)
- src/diagnostic/detail/types/particle.hpp (1 hunks)
- src/diagnostic/diagnostics.hpp (1 hunks)
- src/phare_core.hpp (3 hunks)
- src/python3/patch_level.hpp (6 hunks)
- src/restarts/detail/h5writer.hpp (1 hunks)
- tests/amr/models/test_models.cpp (3 hunks)
- tests/amr/resources_manager/test_resources_manager.cpp (3 hunks)
- tests/core/data/electrons/test_electrons.cpp (6 hunks)
- tests/core/data/ion_population/test_ion_population.cpp (3 hunks)
- tests/core/data/ions/test_ions.cpp (2 hunks)
- tests/core/numerics/ion_updater/test_updater.cpp (6 hunks)
- tests/diagnostic/test_diagnostics.hpp (2 hunks)
Files skipped from review due to trivial changes (2)
- src/core/CMakeLists.txt
- src/core/data/grid/gridlayoutdefs.hpp
Additional comments: 79
tests/amr/models/test_models.cpp (1)
- 33-47: The introduction of
SymTensorField1Dand its integration intoIonsPop1DandIons1Dis correctly done. TheElectronsclass now correctly takesIons1Das a template parameter. These changes indicate the introduction of support for symmetric tensor fields and their integration into the particle and electromagnetic models.src/diagnostic/diagnostics.hpp (1)
- 55-61: Ensure that the
h5::H5Writer<ModelView_t>class has the same interface as the previoush5::Writer<ModelView_t>class, or that all uses ofWriter_thave been updated to match the new interface.src/diagnostic/detail/types/electromag.hpp (2)
69-75: Ensure that the change from
core::Components::componentMaptocore::Components::componentMap()is intentional and correct. IfcomponentMapwas a data member, this change could break the code. If it was a function, this change is correct.104-110: Similar to the previous comment, ensure that the change from
core::Components::componentMaptocore::Components::componentMap()is intentional and correct.src/restarts/detail/h5writer.hpp (5)
1-2: Ensure that the header guard is unique across the codebase.
1-10: > Note: This review was outside of the patch, but no patch was found that overlapped with it. Original lines [14-22]
Ensure that the
Writerconstructor andmake_uniquefunction are being called with the correct arguments in the codebase.
- 1-10: > Note: This review was outside of the patch, but no patch was found that overlapped with it. Original lines [24-40]
Ensure that the
dumpfunction is being called with the correct arguments in the codebase. Also, verify that the file paths and data being written are correct.
- 1-10: > Note: This review was outside of the patch, but no patch was found that overlapped with it. Original lines [42-44]
Ensure that the
modelViewfunction is being called correctly in the codebase.
- 1-10: > Note: This review was outside of the patch, but no patch was found that overlapped with it. Original lines [46-48]
Ensure that the private members
path_andmodelView_are being used correctly in the codebase.tests/diagnostic/test_diagnostics.hpp (2)
36-41: The change from
core::Components::componentMaptocore::Components::componentMap()is correct ifcomponentMapis a function. Ensure that this change is reflected throughout the codebase.45-51: The change from
Writer<ModelView_t>toH5Writer<ModelView_t>is correct ifH5Writeris the new class replacingWriter. Ensure that all instances ofWriterhave been replaced withH5Writerthroughout the codebase.src/phare_core.hpp (2)
33-45: The addition of
SymTensorField_tand the modification ofIonPopulation_tare consistent with the pull request summary. Ensure that these changes are propagated throughout the codebase where these types are used.49-58: The
IonPopulation_ttype alias has been updated to includeSymTensorField_tas a template parameter. This change should be reflected in all instances whereIonPopulation_tis used.src/core/hybrid/hybrid_quantities.hpp (1)
- 13-38: The addition of new enum values for momentum tensor components is a good way to handle the new tensor quantities. Ensure that these new enum values are used correctly throughout the codebase.
tests/core/data/ion_population/test_ion_population.cpp (2)
29-37: The
DummyTensorFieldstruct is introduced here. Ensure that theisUsableandisSettablemethods are correctly implemented according to the intended behavior of the tensor field.60-61: The
IonPopulationstruct is now instantiated withDummyTensorField. Make sure that all methods and properties ofIonPopulationthat interact with the tensor field are compatible withDummyTensorField.src/core/numerics/interpolator/interpolator.hpp (4)
276-295: The
ParticleToMeshclass has been updated to use a generic function objectfuncfor calculating the deposit. This change increases the flexibility of the class, allowing it to handle different types of deposits.301-316: > Note: This review was outside of the patch, so it was mapped to the patch with the greatest overlap. Original lines [301-325]
The
ParticleToMeshclass has been updated to use a generic function objectfuncfor calculating the deposit. This change increases the flexibility of the class, allowing it to handle different types of deposits.
- 332-347: > Note: This review was outside of the patch, so it was mapped to the patch with the greatest overlap. Original lines [332-359]
The
ParticleToMeshclass has been updated to use a generic function objectfuncfor calculating the deposit. This change increases the flexibility of the class, allowing it to handle different types of deposits.
- 493-510: > Note: This review was outside of the patch, so it was mapped to the patch with the greatest overlap. Original lines [472-510]
The
Interpolatorclass has been updated to use theParticleToMeshclass for calculating the deposit for each component of the flux. This change simplifies the code and improves its maintainability.src/diagnostic/detail/types/particle.hpp (1)
- 4-11: The new includes seem to be in order and relevant to the changes described in the summary. Ensure that these new dependencies are properly managed in the build system.
tests/amr/resources_manager/test_resources_manager.cpp (2)
1-9: The new includes for
vecfield.hppandtensorfield.hppare correctly placed and necessary for the newTensorFieldclass and its aliasSymTensorField.29-38: The type definitions have been updated correctly to include the new
SymTensorField1Das a template parameter forIonPopulation1DandIons1D. TheElectrons1Dtype now correctly usesIons1Das a template parameter.tests/core/data/ions/test_ions.cpp (1)
- 34-43: The changes in the type aliases and class templates look good and are in line with the changes described in the pull request summary. Ensure that these changes are reflected throughout the codebase.
src/core/data/grid/gridlayoutimplyee.hpp (2)
94-112: The new arrays for Mxx, Mxy, Mxz, Myy, Myz, and Mzz are initialized with
data.primalfor all components. Ensure that this is the intended behavior. Also, verify that thehybridQtyCenteringarray is updated correctly with these new quantities.311-334: The switch cases for Mxx, Mxy, Mxz, Myy, Myz, and Mzz have been added correctly. However, ensure that the indices
gridData_.iMxx,gridData_.iMxy,gridData_.iMxz,gridData_.iMyy,gridData_.iMyz, andgridData_.iMzzare defined and initialized properly elsewhere in the code.src/core/data/vecfield/vecfield.hpp (3)
17-18: The
VecFieldclass has been replaced with aTensorFieldalias. Ensure that all instances ofVecFieldin the codebase have been updated to reflect this change.29-34: > Note: This review was outside of the patch, so it was mapped to the patch with the greatest overlap. Original lines [20-31]
The
averagefunction now operates onTensorFieldinstances. Ensure that the function is called correctly throughout the codebase.
- 32-34: The
VecFieldNamesstruct remains unchanged. It seems to be used for naming conventions in theVecFieldclass. SinceVecFieldis now an alias forTensorField, ensure that this struct is still relevant and used correctly.src/core/data/tensorfield/tensorfield.hpp (1)
- 1-290: The
TensorFieldclass template and its aliasSymTensorFieldare well-structured and follow good practices. The class provides a clear interface for manipulating tensor fields and includes appropriate error handling. The use ofconstexprfor compile-time calculations andstd::arrayfor fixed-size arrays are good choices for performance and safety. The use ofstd::unordered_mapfor mapping component names to indices is a good choice for efficient lookups. The use ofstd::runtime_errorfor error handling is appropriate. The code is well-documented with comments explaining the purpose of each function. The use ofdeletefor the copy constructor and copy assignment operator prevents copying ofTensorFieldobjects, which is a good practice for classes that manage resources. The use ofdefaultfor the move constructor and move assignment operator enables efficient moving ofTensorFieldobjects. The use ofNO_DISCARDattribute is a good practice to warn users when the return value of a function is discarded. The use ofstd::forward_as_tupleandstd::make_index_sequencein thecomponentsfunction is a good practice for creating tuples of references to the components. The use ofstd::beginandstd::endin thebeginandendfunctions is a good practice for providing range-based for loop support. ThecopyDatafunction checks the usability of both the source and destination tensor fields before copying data, which is a good practice for error handling. ThesetBufferfunction checks the validity of the buffer name before setting the buffer, which is a good practice for error handling. TheisUsableandisSettablefunctions check the usability and setability of the tensor field, which is a good practice for error handling. ThegetComponentandgetComponentNamefunctions check the usability of the tensor field before returning the component or component name, which is a good practice for error handling. Thezerofunction checks the usability of the tensor field before zeroing it, which is a good practice for error handling.src/core/data/vecfield/vecfield_component.hpp (3)
12-12: New components have been added to the
Componentenum. Ensure that these are handled correctly throughout the codebase.16-25: The
componentMapfunction now returns different maps based on the rank. This is a good approach to handle different types of components.35-41: The
checkfunction has been updated to include checks for the new components. This is a necessary change given the addition of new components.src/core/data/ions/ions.hpp (10)
7-11: The new includes are necessary for the changes in this file. Ensure that they are not duplicated elsewhere in the code.
28-34: The new
tensorfield_typeis correctly defined and used.40-46: The
momentumTensor_member is correctly initialized in the constructor.91-96: The
momentumTensor()member functions are correctly defined.143-161: The
computeFullMomentumTensor()function correctly computes the full momentum tensor by summing over all populations. Ensure that themomentumTensor()of each population is correctly computed before this function is called.170-179: The
isUsable()function correctly checks the usability of themomentumTensor_.182-192: The
isSettable()function correctly checks the settable status of themomentumTensor_.204-210: The
MomentPropertiestype has been correctly changed fromstd::vectortostd::array. Ensure that all uses of this type in the codebase have been updated accordingly.234-238: The
getCompileTimeResourcesUserList()function correctly includes themomentumTensor_.259-264: The
momentumTensor_member is correctly declared and initialized.src/diagnostic/detail/h5writer.hpp (10)
23-48: The class
H5Writeris introduced to replace theWriterclass. Ensure that all instances of theWriterclass have been replaced withH5Writerthroughout the codebase.54-64: The constructor of
H5Writeris updated. Ensure that the constructor is called correctly with the new parameters in all instances.82-86: The method
getDiagnosticWriterForTypeis introduced. Ensure that this method is used correctly in the codebase.142-148: The method
writeVecFieldAsDatasetis introduced. Ensure that this method is used correctly in the codebase.162-168: The
typeWriters_member variable is introduced and initialized. Ensure that this variable is used correctly in the codebase.182-185: The copy and move constructors and assignment operators are deleted to prevent shallow copying. This is a good practice for classes that manage resources.
210-215: The
dumpmethod is updated. Ensure that this method is called correctly with the new parameters in all instances.237-243: The
dump_levelmethod is introduced. Ensure that this method is used correctly in the codebase.255-268: > Note: This review was outside of the patch, so it was mapped to the patch with the greatest overlap. Original lines [257-278]
The
initializeDatasets_method is introduced. Ensure that this method is used correctly in the codebase.
- 297-301: The
writeDatasets_method is introduced. Ensure that this method is used correctly in the codebase.src/core/data/ions/ion_population/ion_population.hpp (8)
21-24: The new template parameter
TensorFieldhas been added to theIonPopulationclass. Ensure that all instances of this class have been updated accordingly.33-40: The constructor now initializes a new member
momentumTensor_. Ensure that theinitializerobject passed to the constructor contains the necessary data for this member.54-63: The
isUsableandisSettablemethods now also check the usability and setability ofmomentumTensor_. This is a logical change that should not cause any issues if themomentumTensor_is properly initialized.191-195: New getter methods for
momentumTensor_have been added. These methods should work as expected ifmomentumTensor_is properly initialized.208-214: The type of
MomentPropertieshas been changed fromstd::vectortostd::array. This change should not cause any issues as long as the size of the array is correctly managed.226-231: The type of
ParticlePropertieshas been changed fromstd::vectortostd::array. This change should not cause any issues as long as the size of the array is correctly managed.260-264: The return type of
getCompileTimeResourcesUserListhas been changed to includemomentumTensor_. Ensure that all calls to this method have been updated accordingly.283-289: The new member
momentumTensor_has been added. Ensure that this member is properly initialized and managed throughout the class.src/python3/patch_level.hpp (6)
79-85: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.97-103: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.162-168: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.180-186: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.199-205: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.218-224: The
componentMapfunction is now being called from thecore::Componentsnamespace instead of being a member function. Ensure that this change is consistent across the codebase and that the new function provides the same functionality as the old one.tests/core/numerics/ion_updater/test_updater.cpp (6)
241-258: The addition of momentum tensor fields for the
IonsBuffersstruct is consistent with the pull request summary. Ensure that these fields are used appropriately in the rest of the codebase.298-334: The initialization of the momentum tensor fields in the constructor is correct. The use of
layout.allocSizeensures that the fields are allocated with the correct size.381-418: The copy constructor correctly initializes the momentum tensor fields. It's good that the code is consistent with the original constructor.
467-473: The setting of the momentum tensor fields for the
Ionsobject is correct. This is consistent with the changes made to theIonsBuffersstruct.482-487: The setting of the momentum tensor fields for the proton population is correct. This is consistent with the changes made to the
IonsBuffersstruct.496-501: The setting of the momentum tensor fields for the alpha population is correct. This is consistent with the changes made to the
IonsBuffersstruct.tests/core/data/electrons/test_electrons.cpp (6)
3-7: The new include statement for "core/data/tensorfield/tensorfield.hpp" is added. Ensure that the path is correct and the file exists.
129-134: The new type aliases
SymTensorFieldNDandIonPopulationNDare introduced. Make sure they are used correctly throughout the code.152-163: New fields for the components of the momentum tensor are declared. Ensure that they are initialized before use.
191-214: The fields for the components of the momentum tensor are allocated sizes. Ensure that the sizes are correct and that there is enough memory available.
235-241: The buffers for the components of the momentum tensor are set. Ensure that the buffers are correctly initialized and that the keys used for setting the buffers are correct.
250-255: The buffers for the components of the protons' momentum tensor are set. Ensure that the buffers are correctly initialized and that the keys used for setting the buffers are correct.
| @@ -9,7 +9,7 @@ | |||
| #include "core/data/vecfield/vecfield.hpp" | |||
| #include "core/hybrid/hybrid_quantities.hpp" | |||
|
|
|||
|
|
|||
| #include "core/data/tensorfield//tensorfield.hpp" | |||
There was a problem hiding this comment.
There is a double slash in the include path. This should be corrected to a single slash.
- #include "core/data/tensorfield//tensorfield.hpp"
+ #include "core/data/tensorfield/tensorfield.hpp"Commitable suggestion
[!IMPORTANT]
Carefully review the code before committing. Make sure it correctly replaces the highlighted code, has no missing lines and indentaion issues.
| #include "core/data/tensorfield//tensorfield.hpp" | |
| #include "core/data/tensorfield/tensorfield.hpp" |
| template<std::size_t rank, typename = std::enable_if_t<rank == 1 or rank == 2, void>> | ||
| using TensorType = std::conditional_t<rank == 1, Vector, Tensor>; |
There was a problem hiding this comment.
The TensorType template alias is a good way to handle different ranks of tensors. However, it currently only supports ranks 1 and 2. If you plan to support other ranks in the future, consider making this more flexible.
| static constexpr std::array<Scalar, 6> componentsQuantities(Tensor qty) | ||
| { | ||
| // no condition, for now there's only then momentum tensor M | ||
| return {{Scalar::Mxx, Scalar::Mxy, Scalar::Mxz, Scalar::Myy, Scalar::Myz, Scalar::Mzz}}; | ||
| } |
There was a problem hiding this comment.
The componentsQuantities function for Tensor quantities currently only supports the momentum tensor M. If you plan to support other tensor quantities in the future, consider making this more flexible.
| * this return 0, meaning that this operator | ||
| * have the most priority | ||
| */ | ||
| NO_DISCARD int getOperatorPriority() const override { return 0; } | ||
| int getOperatorPriority() const override { return 0; } |
There was a problem hiding this comment.
The NO_DISCARD attribute was removed from the getOperatorPriority function. This attribute is used to warn developers when the return value of a function is discarded. If the return value of this function is important and should not be ignored, consider adding the NO_DISCARD attribute back.
| * In our case, we allow a RF up to 10, so having 5 ghost width is sufficient | ||
| * | ||
| */ | ||
| NO_DISCARD SAMRAI::hier::IntVector | ||
| getStencilWidth(SAMRAI::tbox::Dimension const& dim) const override | ||
| SAMRAI::hier::IntVector getStencilWidth(SAMRAI::tbox::Dimension const& dim) const override | ||
| { | ||
| return SAMRAI::hier::IntVector{dim, 2}; | ||
| } |
There was a problem hiding this comment.
The NO_DISCARD attribute was removed from the getStencilWidth function and the return type was changed. If the return value of this function is important and should not be ignored, consider adding the NO_DISCARD attribute back. Also, ensure that the change in return type does not affect the rest of the codebase.
| }; | ||
|
|
||
| auto infoVF = [&](auto& vecF, std::string name, auto& attr) { | ||
| for (auto const& [id, type] : core::Components::componentMap) | ||
| for (auto const& [id, type] : core::Components::componentMap<1>()) | ||
| infoDS(vecF.getComponent(type), name + "_" + id, attr); | ||
| }; | ||
|
|
||
| auto infoTF = [&](auto& tensorF, std::string name, auto& attr) { | ||
| for (auto const& [id, type] : core::Components::componentMap<2>()) | ||
| infoDS(tensorF.getComponent(type), name + "_" + id, attr); | ||
| }; |
There was a problem hiding this comment.
The infoTF function is not used anywhere in the code. If it's not needed, consider removing it.
| // if (isActiveDiag(diagnostic, tree, "momentum_tensor")) | ||
| // infoTF(momentum_tensor_["ion"], "momentum_tensor", patchAttributes[lvlPatchID]["ion"]); |
There was a problem hiding this comment.
The commented-out code should be removed if it's not needed. If it's a placeholder for future implementation, consider adding a TODO comment.
| std::string popId{"fluid_" + pop.name()}; | ||
| std::string tree{"/ions/pop/" + pop.name() + "/"}; | ||
| std::string popPath(path + "pop/" + pop.name() + "/"); | ||
| if (checkActive(tree, "density")) | ||
| if (isActiveDiag(diagnostic, tree, "density")) | ||
| initDS(path, attr[popId], "density", null); | ||
| if (checkActive(tree, "flux")) | ||
| if (isActiveDiag(diagnostic, tree, "flux")) | ||
| initVF(path, attr[popId], "flux", null); | ||
| // if (isActiveDiag(diagnostic, tree, "momentum_tensor")) | ||
| // initTF(path, attr[popId], "momentum_tensor", null); | ||
| } | ||
|
|
||
| std::string tree{"/ions/"}; | ||
| if (checkActive(tree, "density")) | ||
| if (isActiveDiag(diagnostic, tree, "density")) | ||
| initDS(path, attr["ion"], "density", null); | ||
| if (checkActive(tree, "bulkVelocity")) | ||
| if (isActiveDiag(diagnostic, tree, "bulkVelocity")) | ||
| initVF(path, attr["ion"], "bulkVelocity", null); | ||
| // if (isActiveDiag(diagnostic, tree, "momentum_tensor")) | ||
| // initTF(path, attr["ion"], "momentum_tensor", null); | ||
| }; |
There was a problem hiding this comment.
The initPatch function contains commented-out code at lines 213-214 and 222-223. If it's not needed, consider removing it. If it's a placeholder for future implementation, consider adding a TODO comment.
| // if (isActiveDiag(diagnostic, tree, "momentum_tensor")) | ||
| // writeTF(path + "momentum_tensor", momentum_tensor_["ion"]); |
There was a problem hiding this comment.
The commented-out code should be removed if it's not needed. If it's a placeholder for future implementation, consider adding a TODO comment.
| inline std::unordered_map<std::string, Component> const Components::vecComponentMap{ | ||
| {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}}; | ||
| inline std::unordered_map<std::string, Component> const Components::tensComponentMap{ | ||
| {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}}; | ||
|
|
There was a problem hiding this comment.
The vecComponentMap and tensComponentMap have been defined but they both map to the same components (X, Y, Z). The tensComponentMap should map to the tensor components (XX, XY, XZ, YY, YZ, ZZ).
- inline std::unordered_map<std::string, Component> const Components::tensComponentMap{
- {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}};
+ inline std::unordered_map<std::string, Component> const Components::tensComponentMap{
+ {"xx", Component::XX}, {"xy", Component::XY}, {"xz", Component::XZ},
+ {"yy", Component::YY}, {"yz", Component::YZ}, {"zz", Component::ZZ}};Commitable suggestion
[!IMPORTANT]
Carefully review the code before committing. Make sure it correctly replaces the highlighted code, has no missing lines and indentaion issues.
| inline std::unordered_map<std::string, Component> const Components::vecComponentMap{ | |
| {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}}; | |
| inline std::unordered_map<std::string, Component> const Components::tensComponentMap{ | |
| {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}}; | |
| inline std::unordered_map<std::string, Component> const Components::vecComponentMap{ | |
| {"x", Component::X}, {"y", Component::Y}, {"z", Component::Z}}; | |
| inline std::unordered_map<std::string, Component> const Components::tensComponentMap{ | |
| {"xx", Component::XX}, {"xy", Component::XY}, {"xz", Component::XZ}, | |
| {"yy", Component::YY}, {"yz", Component::YZ}, {"zz", Component::ZZ}}; |
#752
Computing the pressure is like using the ParticleToMesh except we need to deposit neither 1 nor the velocity, but the velocity squared. The idea thus is to generalize the ParticleToMesh operator so, per particle, to deposit only 1 field at a time, given the field, the particle and a functor that, given the particle would:
for the density, return 1
for the flux, return vx, or vy or vz
for the pressure, return vxvx, vxvy etc.
This is done so that one day if needed to compute something else, like the heat flux (v^3) we can simply add more functors.
Summary by CodeRabbit
TensorFieldclass for representing tensor fields.IonPopulationclass.Interpolatorclass to handle tensor fields.FluidDiagnosticWriterclass.Writerclass toH5Writer.VecFieldclass with aTensorFieldalias.componentMapfunction calls in various files.