A collection of .sh and .py scripts to handle files generated by the OPTIC programs within WIEN2k.
By Levi Tegg
WIEN2k is a collection of programs used to perform electronic structure calculations of solids, using density functional theory. Once a calculation has been completed, the OPTIC routine (consisting of programs lapw2 -fermi, optic, joint, and kram) will generate files containing optical properties data, such as the dielectric function (*.epsilon), or the optical conductivity (*.absorp).
This repository contains scripts which parse some of the files written by WIEN2k and OPTIC, then calculate optical constants such as complex refractive index, electron energy-loss spectra (EELS), and an approximation of the apparent colour.
This repository contains both Python (.py) and bash (.sh) scripts. In general, each script can be installed and used independently of the other scripts within this repository.
All of the scripts were written using Python 3.5. Though each script differs slightly, the main libraries used are numpy, pyplot from matplotlib, glob, and sys. These scripts were written using Spyder 3.1.x, using the Anaconda3 distribution.
The bash scripts are largely handlers for the Python scripts. They should work with any bash-like terminal.
Some of the scripts (e.g., optics.py) should work anywhere. In some cases, a .sh will call a .py script, which is described by a path (e.g., ETEST.sh and ETEST.py). In such cases, inspect and edit the .sh files as necessary.
Most of the python scripts include code to search the present working directory (pwd) for the files it needs. Alternatively, arguments can be passed to the script to tell it exactly what to process.
Calculates optical properties using a *.epsilon file. Calling
$ python3 optics.py
in a directory with one or more *.epsilon files will process all of the *.epsilon files in the pwd. Alternatively, calling
$ python3 optics.py case1.epsilon case2.epsilon case3.epsilon ...
will make optics.py process only those epsilon files passed as arguments.
If 3 columns are found in a *.epsilon file, the script will assume a one-dimensional case. If 5 columns are found, the script will assume the x- and y-directions are equal, and the z-direction is different. If 7 columns are found, the script will assume the x-, y- and z-directions are different. Currently, there is no support for more than 3-directions.
The script will save images and data files with the same case name as the *.epsilon file, in the same directory as the *.epsilon file.
Further information can be found in the script comments.
After optic has finished, calling
$ bands.sh
will grep the *.output2 file and print the number of bands, as well as the Fermi energy, for use in *.injoint.
After joint with option 6 has been run, calling
$ wp.sh
will grep the *.outputjoint and print the plasma frequencies, for use in *.inkram.
In a directory where a non-parallelised lapw1 is running, calling
$ lapw1_check.sh
will report how many k-points have been completed compared to the total number. Useful for checking how much longer a given iteration of lapw1 has remaining. Note that it doesn't (yet) work for parallelised cases, where multiple *.output1 files are generated.
In a directory where a STDOUT file is being generated by WIEN2k and w2web, calling
$ ETEST.sh
will parse the STDOUT file, then call ETEST.py, and produce a plot of the ETEST parameter as a function of the number of cycles. The script then calls Eye of Gnome (eog) to automatically open the image. On systems without eog, the image will still be produced, but the shell script will return an error.
These scripts are designed for a very specific purpose, and so have not been designed with robustness or security in mind. As such, there are probably very many usage scenarios where these scripts produce incorrect results or fail completely. The author accepts no responsibility for results obtained using these scripts, and encourages users to read the scripts in their entirety before use. That said, there are some issues and bugs that I am aware of (and you should be too):
- Accurately interpolating wavelength and reflectivity for the calculation of the apparent colour in
optic.pyrelies on a fine energy spacing (in eV) across the visible energy range. The default values in*.inopare generally sufficient, but check your discretisation to ensure there are no sudden changes in the refractive index. - When 2 or more dimensions are present some of the graphs generated by
optics.pyinclude optical constants for each direction (x, x,z or x,y,z), as well as a weighted sum (x+y+z). This is calculated via xyz = (x + y + z) / 3. Be aware that in the case of EELS data, the true loss spectrum for a polycrystalline sample of an anisotropic material cannot be calculated in this way. ETEST.shandETEST.pyassume that the output from the SCF cycle is being printed to a file calledSTDOUTin the directory of the calculation.w2webdoes this by default, but this is not typical behaviour whenrun_lapwis called from the command line. Useanalyse_scfto parse the*.scffile for other quantities.
I've uploaded the scripts onto GitHub so that they may be shared, repurposed and edited freely. That said, if you find an issue with the scripts that you'd like to discuss or have fixed, you are welcome to submit a pull request or issue.