The CRYSTAL is an ab initio solid state modeling suite employing the Gaussian basis sets in the LCAO framework. The pycrystal Python utilities are good for:
- quick logs parsing, getting the maximum information, and presenting it in a systematic machine-readable way
- preparing and handling the Gaussian LCAO basis sets, based on the EMSL and own CRYSTAL libraries
All the popular versions of the CRYSTAL code are supported (CRYSTAL03, CRYSTAL06, CRYSTAL09, CRYSTAL14, and CRYSTAL17).
The pycrystal was tested on about 20k in-house simulation logs for about 700 distinct materials systems, produced with the different CRYSTAL versions. Its development was initiated in 2009 by Maxim Losev at the quantum chemistry chair, chemistry dept. of St. Petersburg State University (Russia) under supervision of Professor Robert Evarestov.
pip install pycrystal
Parsing is done as follows:
import os, sys
from pprint import pprint
from pycrystal import CRYSTOUT
try:
sys.argv[1] and os.path.exists(sys.argv[1])
except (IndexError, OSError):
sys.exit("USAGE: <script> <file>")
assert CRYSTOUT.acceptable(sys.argv[1])
result = CRYSTOUT(sys.argv[1])
pprint(result.info)Also, for any basis set taken from EMSL in Gaussian'94 format:
import os, sys
from pycrystal import parse_bs
try:
sys.argv[1] and os.path.exists(sys.argv[1])
except (IndexError, OSError):
sys.exit("USAGE: <script> <file>")
content = open(sys.argv[1]).read()
for atom in parse_bs(content):
print(atom.crystal_input())To deal with the CRYSTAL (University of Turin) online BS library:
from pycrystal import download_basis_library
library = download_basis_library()
print(library['Bi'])There is another Python parser ejplugins for CRYSTAL14 and CRYSTAL17 by Chris Sewell (Imperial College London, UK). The comparison was done using cmp_unito_crystal_parsers.py script on the above-mentioned 20k logs, the results are as follows:
- the final total energies and atomic structures are the same in more than 99% cases
pycrystalsupports slightly more CRYSTAL features thanejpluginspycrystalis more lightweight thanejpluginsand has less dependencies- performance is nearly the same
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent