6.3. How to use HORTON to convert between file formats?

As discussed in the Running HORTON as horton-*.py scripts section, the (non-)optional command-line arguments of the conversion scripts can be looked at by the --help flag:

horton-convert.py --help
horton-hdf2csv.py --help

6.3.1. horton-convert.py – Conversion between file formats supported by HORTON

The horton-convert.py script converts between various file formats supported by HORTON. To use it, simply provide the input and output files, as follows:

horton-convert.py input_file output_file

The formats you are converting from/to are deduced from the input/output file name specified. This conversion only works if both file formats are recognized by HORTON, and the required information to write the output file is available in the input file. For more information, please refer to Data file formats (input and output).

This script is typically useful for archiving large amounts of data produced by other codes in text format. The following example converts a *.cube file to HORTON’s internal HDF5 format, and then uses h5repack, one of the HDF5 command-line tools, to compress the binary file.

horton-convert.py density.cube density_cube.h5
h5repack -v -f GZIP=1 density_cube.h5 density_cube_gzip.h5

6.3.2. horton-hdf2csv.py – Conversion of HDF5 files to CSV format

The horton-hdf2csv.py script is used by providing the HDF5 and the CSV file names, respectively, as follows:

horton-hdf2csv.py somefile.h5:path/in/hdf5/file otherfile.csv

For example, if horton-wpart.py was used to run an Extended Hirshfeld partitioning of a wavefunction in gaussian.fchk file format, the command below would convert the results into CSV file format:

horton-hdf2csv.py wpart_he_results.h5 extended_hirshfeld.csv

This script was added for the convenience of those who are not familiar with the HDF5 files. If you are familiar with the HDF5 files, we strongly suggest you to directly process these files with conventional (Python) scripts. This would be far easier than interfacing with the CSV files. The h5py library is a great tool to make such custom scripts.