Improving the coupling between SLOTH and thermodynamic data

[CPLUMECOCQ]

The aim of this problem is to be able to couple abstract thermodynamic data with our phase field solver, to do this the steps are:

• Read/write data -> we have chosen to use the HDF5 format https://www.hdfgroup.org/solutions/hdf5/
• Perform linear/bilinear/trilinear interpolation
• Perform n-linear interpolation

[CPLUMECOCQ]

Here are a few points for discussion @CINTROINI :

Current additions:

• Multilinear interpolator for boost::multiarray structures
• Class for reading/writing HDF5 files

Questions/discussion points:

• Should we make ONE very generic interpolation class, or one class for single-phase thermo, one for two-phase, ...?
• Which nomenclature for parameters?
• What nomenclature for HDF5 files?
• Do we keep the simple interpolation from a .csv for binaries?

note: do not consider the following files (they will be deleted from the merge request):

• kernel/CMakeLists.txt
• kernel/Calphad/GeneralMultiParamsTabulation.hpp
• kernel/sloth.hpp (partial)
• kernel/Calphad/BinaryTabulation.hpp (see questions)