PyBigDFT Development Guide¶
PyBigDFT is the high level python interface which you can use to drive BigDFT calculations. PyBigDFT does not directly interface with any Fortran routines, and instead only interfaces using the file system. When you wish to add new high level features to the code, this can be the ideal entry point.
The shared language between the Fortran code of BigDFT and python is the YAML markup language. Thanks to Futile, BigDFT’s Fortran can still flexibily handle complex input. Python dictionaries can be easily converted into YAML. For this reason, many of the PyBigDFT classes inherit from dictionary so that they can easily be serialized to a yaml format.
When adding new functionality to PyBigDFT, you should document each module according to the Google Doc String style. This will ensure that the API is automatically documented.
Each module of PyBigDFT should include an _example subroutine which demonstrates the functionality of the package. This example should strive for as much code coverage as possible. After a new feature has been created, and added to the _example routine, you can run the regression tests using the following command:
python -m unittest discover -s $PYTHONPATH/BigDFT/ -p "check_examples.py"
This command can be run in any directory after sourcing the BigDFT variables. You will want to do this in a dedicated test directory as the results of each test are written to files in the current directory.
Major functionality should be documented as a tutorial or lesson as a Jupyter notebook. Once a notebook has been created, you should add it to the bigdft-doc project.
Regression for the notebooks.