Skip to content

Comments

Refactoring and replacing of old ad hoc file formats for HDF5 to start aligning with NeXus for the NOMAD Oasis project via the FAIRmat consortium of the German NFDI#2

Open
mkuehbach wants to merge 14 commits intomasterfrom
twod_obc_solver_nexus_io
Open

Refactoring and replacing of old ad hoc file formats for HDF5 to start aligning with NeXus for the NOMAD Oasis project via the FAIRmat consortium of the German NFDI#2
mkuehbach wants to merge 14 commits intomasterfrom
twod_obc_solver_nexus_io

Conversation

@mkuehbach
Copy link
Owner

No description provided.

atomprobe-tc and others added 14 commits March 11, 2024 15:43
… this was successful, so we can export quantities now easier with HDF5, next step implement export of quantities so far stored in Texture ori files do so for now as flat fields, thereafter export polygonal contour of grains such that they can be rendered with XDMF/Paraview
…imple NeXus-like HDF5 file with compression, next step contour geometry and XDMF to Paraview
…MF Polygon instances, working, current limitations of this implementation: i) polygon faces cannot currently be colored based on properties at least when using paraview as edge-centered XDMF attribute data are at least for paraview v5.12.0rc1 not supported, ii) no incremental writing of the vertex coordinate and xdmf information which may be prohibitive for large simulations run on computers with insufficient main memory, iii) vertices cannot be reduced to the set of unique to get the compromise of being able to filter individual grains using the threshold filter by grain id and have their contour remaining closed, iv) observation with relevance for research: this vector based visualization (in contrast to earlier gnuplots) is more performant but shows that gaps exists between grains. These have been mentioned by Christian and are numerically closed during the repartition step of the level set functions, see his PhD for details but as currently the size of the grains is also computed based on the getRawBoundary contour (the inscribed varea of respective polygons) this can explain that the total area of the grain ensemble does not add up completely to 1 but close to by +- 0.1 %.
…cture_generator that works with NeXus/HDF5, next steps i) change loading of data in twod_obc_solver, ii) refactor input for twod_obc_solver, iii) use proper NeXus/HDF5
…be tested, maybe this commit needs to be reverted!
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants