def __del__(self):
"""
Garbage collection to destroy the neighbor list automatically.
"""
if self.neigh is not None:
nl.clean(self.neigh)
self.neigh = None
def test_main():
# create contributing atoms
alat = 2.46
d = 3.35
cell, contrib_coords, contrib_species = create_graphite_unit_cell(alat, d)
# create padding atoms
cutoffs = np.array([d + 0.01, d + 0.02], dtype=np.double)
influence_dist = cutoffs[1]
pbc = np.array([1, 1, 1], dtype=np.intc)
out = nl.create_paddings(influence_dist, cell, pbc,
contrib_coords, contrib_species)
pad_coords, pad_species, pad_image, error = out
check_error(error, 'nl.create_padding')
assert pad_coords.shape == (96, 3)
# print('pad_coords is of shape:', pad_coords.shape)
coords = np.concatenate((contrib_coords, pad_coords))
coords = np.asarray(coords, dtype=np.double)
species = np.concatenate((contrib_species, pad_species))
species = np.asarray(species, dtype=np.intc)
fname = 'atoms.xyz'
write_XYZ(fname, cell, species, coords)
# flag to indicate wheter create neighbor list for an atom
n_pad = pad_coords.shape[0]
n_contrib = contrib_coords.shape[0]
need_neigh = np.concatenate((np.ones(n_contrib), np.zeros(n_pad)))
need_neigh = np.asarray(need_neigh, dtype=np.intc)
# create neighbor list
neigh = nl.initialize()
error = nl.build(neigh, coords, influence_dist, cutoffs, need_neigh)
check_error(error, 'nl.build')
# build again (it will automatically empty previous neigh list)
error = nl.build(neigh, coords, influence_dist, cutoffs, need_neigh)
check_error(error, 'nl.build')
# test get neigh function
neigh_list_index = 0
particle = 1
num_neigh, neighbors, error = nl.get_neigh(
neigh, cutoffs, neigh_list_index, particle)
check_error(error, 'nl.get_neigh')
assert num_neigh == 14
# print('Atom 1 has {} neighbors:'.format(num_neigh), end=' ')
# for i in neighbors:
# print(i, end=' ')
neigh_list_index = 1
particle = 4
num_neigh, neighbors, error = nl.get_neigh(
neigh, cutoffs, neigh_list_index, particle)
check_error(error, 'nl.get_neigh')
assert num_neigh == 0
# expect error message from this
# neigh_list_index = 1
# particle = n_contrib + n_pad
# num_neigh, neighbors, error = nl.get_neigh(neigh, cutoffs, neigh_list_index, particle)
# assert error == 1
# delete neighbor list
nl.clean(neigh)
# remove the created file
try:
os.remove(fname)
os.remove('kim.log')
except:
pass
def clean(self):
neighlist.clean(self.neigh)
def test_main():
modelname = 'ex_model_Ar_P_Morse_07C'
print()
print('=' * 80)
print('Matching results for KIM model:', modelname)
print()
# create model
requestedUnitsAccepted, kim_model, error = kimpy.model.create(
kimpy.numbering.zeroBased,
kimpy.length_unit.A,
kimpy.energy_unit.eV,
kimpy.charge_unit.e,
kimpy.temperature_unit.K,
kimpy.time_unit.ps,
modelname,
)
check_error(error, 'kimpy.model.create')
if not requestedUnitsAccepted:
report_error('requested units not accepted in kimpy.model.create')
# units
l_unit, e_unit, c_unit, te_unit, ti_unit = kim_model.get_units()
check_error(error, 'kim_model.get_units')
print('Length unit is:', str(l_unit))
print('Energy unit is:', str(e_unit))
print('Charge unit is:', str(c_unit))
print('Temperature unit is:', str(te_unit))
print('Time unit is:', str(ti_unit))
print()
# create compute arguments
compute_arguments, error = kim_model.compute_arguments_create()
check_error(error, 'kim_model.compute_arguments_create')
# check compute arguments
num_compute_arguments = (
kimpy.compute_argument_name.get_number_of_compute_argument_names())
print('Number of compute_arguments:', num_compute_arguments)
for i in range(num_compute_arguments):
name, error = kimpy.compute_argument_name.get_compute_argument_name(i)
check_error(error, 'kim_model.get_compute_argument_name')
dtype, error = \
kimpy.compute_argument_name.get_compute_argument_data_type(name)
check_error(error, 'kim_model.get_compute_argument_data_type')
support_status, error = \
compute_arguments.get_argument_support_status(name)
check_error(error, 'compute_argument.get_argument_support_status')
n_space_1 = 21 - len(str(name))
n_space_2 = 7 - len(str(dtype))
print('Compute Argument name "{}" '.format(name) + ' ' * n_space_1 +
'is of type "{}" '.format(dtype) + ' ' * n_space_2 +
'and has support status "{}".'.format(support_status))
# can only handle energy and force as a required arg
if support_status == kimpy.support_status.required:
if name not in (kimpy.compute_argument_name.partialEnergy,
kimpy.compute_argument_name.partialForces):
report_error('Unsupported required ComputeArgument')
# must have energy and forces
if name in (kimpy.compute_argument_name.partialEnergy,
kimpy.compute_argument_name.partialForces):
if support_status not in (kimpy.support_status.required,
kimpy.support_status.optional):
report_error('Energy or forces not available')
print()
# check compute callbacks
num_callbacks = \
kimpy.compute_callback_name.get_number_of_compute_callback_names()
print('Number of callbacks:', num_callbacks)
for i in range(num_callbacks):
name, error = kimpy.compute_callback_name.get_compute_callback_name(i)
check_error(error, 'kim_model.get_compute_callback_name')
support_status, error = \
compute_arguments.get_callback_support_status(name)
check_error(error, 'compute_argument.get_callback_support_status')
n_space = 18 - len(str(name))
print('Compute callback "{}"'.format(name) + ' ' * n_space +
'has support status "{}".'.format(support_status))
# cannot handle any "required" callbacks
if support_status == kimpy.support_status.required:
report_error('Unsupported required ComputeCallback')
print()
# parameter
num_params = kim_model.get_number_of_parameters()
print('Number of parameters:', num_params)
print()
for i in range(num_params):
out = kim_model.get_parameter_metadata(i)
dtype, extent, name, description, error = out
check_error(error, 'kim_model.get_parameter_metadata')
print('Parameter No.', i)
print(' data type:', dtype)
print(' extent:', extent)
print(' name:', name)
print(' description:', description)
print()
# register argument
argon = create_fcc_argon()
coords = np.asarray(argon.get_positions(), dtype=np.double)
N = coords.shape[0]
print('Number of particles:', N)
forces = np.zeros((N, 3), dtype=np.double)
energy = np.array([0.0], dtype=np.double)
num_particles = np.array([N], dtype=np.intc)
species_code = np.zeros(num_particles, dtype=np.intc)
particle_contributing = np.zeros(num_particles, dtype=np.intc)
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.numberOfParticles, num_particles)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.particleSpeciesCodes, species_code)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.particleContributing,
particle_contributing)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.coordinates, coords)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.partialEnergy, energy)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
error = compute_arguments.set_argument_pointer(
kimpy.compute_argument_name.partialForces, forces)
check_error(error, 'kimpy.compute_argument.set_argument_pointer')
# create neighbor list
neigh = nl.initialize()
# register get neigh callback
error = compute_arguments.set_callback_pointer(
kimpy.compute_callback_name.GetNeighborList, nl.get_neigh_kim(), neigh)
check_error(error, 'kimpy.compute_argument.set_callback_pointer')
# influence distance and cutoff of model
model_influence_dist = kim_model.get_influence_distance()
out = kim_model.get_neighbor_list_cutoffs_and_hints()
model_cutoffs, padding_not_require_neigh_hints = out
print('Model influence distance:', model_influence_dist)
print('Model cutoffs:', model_cutoffs)
print('Model padding neighbors hints:', padding_not_require_neigh_hints)
print()
# species support and code
species_support, code, error = \
kim_model.get_species_support_and_code(kimpy.species_name.Ar)
check_error(error or not species_support,
'kim_model.get_species_support_and_code')
print('Species Ar is supported and its code is:', code)
print()
# setup particle species
species_code[:] = code
# setup particleContributing
particle_contributing[:] = 1
# setup neighbor list
need_neigh = np.ones(N, dtype='intc')
# compute energy and force for different structures
alat = 5.26
min_alat = 0.8 * 5.26
max_alat = 1.2 * 5.26
inc_alat = 0.025 * alat
all_alat = np.arange(min_alat, max_alat, inc_alat)
print('=' * 80)
print('Result for KIM model:', modelname)
print()
print(' energy force norm lattice spacing')
print()
for a in all_alat:
argon = create_fcc_argon(a)
# NOTE cannot change coords address
np.copyto(coords, argon.get_positions())
error = nl.build(neigh, coords, model_influence_dist, model_cutoffs,
need_neigh)
check_error(error, 'nl.build')
error = kim_model.compute(compute_arguments)
print('{:18.10e} {:18.10e} {:18.10e}'.format(energy[0],
np.linalg.norm(forces),
a))
# destory neighbor list
nl.clean(neigh)
# destory compute arguments
error = kim_model.compute_arguments_destroy(compute_arguments)
check_error(error, 'kim_model.compute_arguments_destroy')
# destory model
kimpy.model.destroy(kim_model)