Change log#

v2.1.3#

Change of np.arange to np.linspace to avoid e_grid arrays with sizes different from NEDOS. (see the warning on numpy.arange documentation)

Add bg option to TASGenerator.from_vasp_outputs() to allow scissor shifting of the bandgap to match experiment.
Implement vectorised band filtering functions, massive speedup in processing times.
Implement memory sharing for large arrays within multiprocessing.

Reduce the sum over band transitions to the energy mesh (min(-6, -energy_max), max(6, energy_max)) to make parsing of VASP optics more efficient.
Switch to semantic versioning.

Add multiprocessing functionality to dramatically speed up parsing of large kpoint sets / large numbers of electronic bands, and add example notebook and HPC script for this.

Add functionality to parse orbital derivatives from VASP optics calculations, to incorporate oscillator strengths in the prediction of TAS spectra.
Add functionality to save and load Tas objects to json file, to save time on re-analysing later on.
Update plotting to avoid overlapping lines, ensure colour-consistency when the number of transitions changes, remove redundant transitions from the legend, pass kwargs to plt.legend() etc.
Add the from_vasp_outputs() class method for TASGenerator() to directly generate from VASP outputs (streamlines this process).
Add yaxis = “jdos_diff”, “tas_absorption_only” and “alpha” plotting options.
Add tests and examples for all the above changes.
temp and conc are passed through from the initial user specification, to avoid confusion if changing later in plotting functions and expecting TAS to be re-calculated.
General clean-up, avoidance of unnecessary warnings, addition of error catches, docstrings expansion…