Configure mtex¶
The pynxtools-microstructure plugin for pynxtools connects the Matlab software mtex-toolbox/mtex for orientation and texture analysis to pynxtools and thus optionally to the NOMAD research data management system. The software pynxtools-microstructure provides functionalities to export Matlab class objects from MTex to an HDF5 files that follows the NeXus data model, specifically contributed base classes such as NXmicrostructure, NXem_ebsd, and the NXem application definition. We connect MTex to pynxtools-microstructure via a GitHub submodule.
Instead of connecting to the mtex-toolbox/mtex GitHub repository directly, we opted to work with a FAIRmat/NOMAD-specific fork. This fork we update when necessary, tracking the develop branch of MTex and the FAIRmat-NFDI/mtex forks respectively by default. The fork is imported with a git submodule.
pynxtools-microstructure provides two key software tools: The first, located in src/pynxtools_microstructure/matlab, yields MTex-specific MATLAB code for exporting Matlab class objects to a NeXus/HDF5 file. This tool implements two functionalities: src/pynxtools_microstructure/matlab/mtex/extern/hdfutils, Matlab scripts to write HDF5 files (with file name ending .mtex.h5) and src/pynxtools_microstructure/matlab/mtex/extern/scripts/Markus, Matlab scripts that implement the respective MTex calls. The second key software tool is located in src/pynxtools_microstructure. This tool is Python code that implements a reader for the above-mentioned NeXus/HDF5 files from MTex using the user scripts from the first key software tool. The second key software tool uses the dataconverter of pynxtools to convert this NeXus/HDF5 file. The second tool assures that the NeXus/HDF5 follows typical conventions used by other pynxtools plugins and implements a few other functionalities which the above-mentioned hdfutils Matlab scripts does not implement yet.
The main reason for this double structure is that it is tricky and not generally considered robust how to force Matlab into using a specific version of its HDF5 library, specifically a version that offers multithreaded compression capabilities. Instead, Mathworks ships each Matlab version with a specific version of HDF5. These versions are typically not compiled, though, against advanced compression libraries like blosc2. Therefore, using HDF5 in Matlab in most practical cases and simple cases means working with the sequential compression deflate compression algorithm. Enabling a usage of advanced compression algorithms was another motivation connect to MTex via a fork MTex.
How to update FAIRmat-NFDI/mtex when changes occur on the official mtex-toolbox side¶
Exemplified for updating the FAIRmat-NFDI fork's develop branch with the official MTex repository's develop branch (our current choice).
git clone https://github.com/FAIRmat-NFDI/mtex.git
cd mtex
git checkout develop
git remote add upstream https://github.com/mtex-toolbox/mtex.git
# git remote -v
git fetch upstream
git merge upstream/develop
git push
How to update pynxtools-microstructure to use a specific commit of the FAIRmat-NFDI/mtex fork¶
git clone https://github.com/FAIRmat-NFDI/pynxtools-microstructure.git
git submodule sync --recursive
git submodule update --init --recursive --jobs=4
cd pynxtools-microstructure/src/pynxtools_microstructure/mtex
git fetch
git pull
# get latest commit id via e.g. `git log`
cd ../../../
./scripts/mtex.sh checkout <COMMITID>
# followed by the typical git add, commit pattern