diff --git a/python/preprocessing/README.md b/python/preprocessing/README.md index c916a712d57f29fed8a4f47c75366841519a3a1e..25b9c764a9453e5e8b8b254f436c9f2b70df8a4e 100644 --- a/python/preprocessing/README.md +++ b/python/preprocessing/README.md @@ -10,37 +10,42 @@ some of the SHEMAT-Suite input arrays into hdf5-files. ## HDF5 Input ## -The option to use HDF5 input instead of serial ASCII input is now -available in branch `master` and Branches `mpi-sm-hdf5`. The -Python-script `convert_to_hdf5.py` converts parts of the standard -ASCII input file into and hdf5-file called `<input>_conv.h5`. Parts -which are not converted to hdf5-arrays are stored in a new ASCII-file -called `<input>_conv`. Due to changes in the SHEMAT-Suite reading -routines, most of the input data is then read from the HDF5-file which -is considerably faster than reading from the ASCII-file and has -advantages of parallel I/O. +# Input File Conversion to HDF5-format -How to use it: +The user input for SHEMAT-Suite is provided in an ASCII-file. For reducing the +reading time of larger input files, SHEMAT-Suite provides the option to read +HDF5 input instead. Theresore, we provide a Python-script for converting the +ASCII input file, which is created by the user, into an HDF5 input file. -1.) Copy the python code `convert_to_hdf5.py` from the code directory -to your model directory. +The script `convert_to_hdf5.py` converts parts of the standard ASCII input file +into and HDF5-file called `<input>_conv.h5`. Those parts which are not converted +to hdf5-arrays are stored in a new ASCII-file called `<input>_conv`. +Most of the input data, especially large arrays, is then read from the +HDF5-file, which is considerably faster than reading from the ASCII-file. -2.) For using the converter, the Python module has to be loaded. +**How to use it:** -3.) Conversion of standard SHEMAT-Suite input file: +0.) Prerequisite: Python + +1.) Download the python scipt `convert_to_hdf5.py` +from https://git.rwth-aachen.de/SHEMAT-Suite/shemat-suite_scripts-open + +2.) Conversion of standard SHEMAT-Suite input file: `./convert_to_hdf5.py <inputfile>` -4.) The file shemade.job has to contain the name of the newly -generated ASCII-file `<inputfile>_conv` +3.) Write the name of the newly generated ASCII-file `<inputfile>_conv` +into the file `shemade.job `. + +4.) Execute SHEMAT-Suite. -5.) Execute SHEMAT-Suite. +**Important note:** -Important note: The input of multiple values for e.g. boundary -conditions either has to be written in one line or the number of -entries has to be defined by using the keyword records. Otherwise the -converter will give an error! +The input of multiple values for e.g. boundary +conditions either has to be written in one line or the number of entries +has to be defined by using the keyword records. Otherwise the converter +will give an error! -Please test the HDF5 input option with your models and report any -difficulties like the one noted above, so that they can be adapted in -the future. +Please test the HDF5 input option with your models and report any +difficulties like the one noted above (e.g. as an issue), +so that they can be adapted in the future. \ No newline at end of file