sparc-dft-api is an ASE based python wrapper for the density functional theory (DFT) code SPARC. This wrapper requires Python3, and is currently in active development and alpha testing, so please use it with caution and report any errors in the "Issues" and/or submit pull requests with patches to fix issues as needed.
The following command should install the package:
pip install git+https://github.com/SPARC-X/sparc-dft-api
prerequisites:
python packages:
scipy
numpy
ase
spglib
software:
SPARC
(https://github.com/SPARC-X/SPARC)
Currently installation is barebones, simply clone this repository (git clone https://github.com/SPARC-X/sparc-dft-api.git) and add the folder to your pythonpath (export PYTHONPATH=[were you cloned the repository]:$PYTHONPATH)
To enable the SPARC calculator to work you need to tell it what command to use to run sparc and where your pseudopotentials are. These should be stored in the environment variables $ASE_SPARC_COMMAND and $SPARC_PSP_PATH respectively. e.g.:
export ASE_SPARC_COMMAND=[command to run SPARC]
export SPARC_PSP_PATH=[location of pseudopotentials]
The ASE_SPARC_COMMAND variable is somewhat tricky to set up. All the information about parallelization should be included where appropriate in the command (i.e. mpirun -np X.) SPARC takes in a flag -name followed by the prefix of the input file names. This wrapper deals with the prefixes completely internally, but the word PREFIX should be put in after the -name flag for SPARC. Here are two examples command if the sparc executable is in your $PATH environment variable:
export ASE_SPARC_COMMAND="mpirun -np $PBS_NP sparc -name PREFIX"
export ASE_SPARC_COMMAND="mpirun -rmk pbs sparc -name PREFIX"
note that the exact word "PREFIX" should be contained in this command.
SPARC_PSP_PATH is a path that points to the location of your pseudopotential files. These must have the naming convention: [element name].pot to be detected. Currently SPARC only uses the psp8 format. Alternatively, rather than setting this environment variable you may pass the information in with the pseudo_dir argument.
Once those environment varibles are in place, the SPARC ASE calculator works like any other ASE calculator. It must be imported, instantiated, and called. There are examples of common operations at the bottom of this README.
The SPARC calculator behaves similarly to other ASE calculators, taking an atoms object to define the system being calculated and a label argument for the filename prefixes. The arguments used to control the SPARC software are identical to those of the SPARC software program. Docuymentation can be found at the link below:
https://github.com/SPARC-X/SPARC/blob/master/doc/Manual.pdf
A mesh spacing or finite difference grid must be defined for the calculator to work. This can be done in one of three ways:
- by inputting the MESH_SPACING argument. This will use SPARC's internal mesh spacing to generate a grid
- by using the
hargument, this will default to converting to MESH_SPACING and utilizing this arguement in SPARC - Inputting the
fd_gridarugment which explicitly defines the finite difference grid. For examplefd_grid=[25,25,25]
#get sparc calculator
from sparc.sparc_core import SPARC
calc = SPARC(h=0.2) # a grid spacing grid must be entered.
#make atoms
from ase.build import bulk
atoms = bulk('Si',cubic=True)
atoms.set_pbc([True] * 3)
atoms.set_calculator(calc)
atoms.get_potential_energy()
from sparc.sparc_core import SPARC
calc = SPARC(h=0.2, RELAX_FLAG=1) # a grid spacing grid must be entered.
#make atoms
from ase.build import molecule
atoms = molecule('H2')
atoms.set_cell([6,6,6])
atoms.center()
atoms.set_pbc([False] * 3 )
atoms.set_calculator(calc)
atoms.get_potential_energy()
from sparc.sparc_core import SPARC
calc = SPARC(h=0.2, RELAX_FLAG=1) # a grid spacing grid must be entered.
#make atoms
from ase.build import molecule
atoms = molecule('H2')
atoms.set_cell([6,6,6])
atoms.center()
atoms.set_pbc([False] * 3 )
atoms.set_calculator(calc)
calc.write_input()