def find_atomic_config_document(**kwargs):
atomic_configs = AtomicConfig.objects(__raw__=kwargs).order_by('-_id').all()
ret = []
for atomic_config in atomic_configs:
# may need 'ase' in here somewhere
atomic_config.file_payload = fsc.retrieve(atomic_config.file_uid)
yield atomic_config
def test_custom():
test_path = os.path.join(BASE_PATH, str(uuid.uuid4()) + '.npy')
dd = np.random.rand(500, 500)
with fs_write.NpyWriter(test_path, resource_kwargs={'mmap_mode': 'r'}) as f:
eid = f.add_data(dd)
with fsr.handler_context({'npy': fs_read.NpyHandler}):
ret = fsc.retrieve(eid)
assert_array_equal(dd, ret)
def find_pdf_data_document(**kwargs):
pdf_data_sets = PDFData.objects(__raw__=kwargs).order_by('-_id').all()
for data_set in pdf_data_sets:
data_set.file_payload = fsc.retrieve(data_set.file_uid)
yield data_set
def find_atomic_config_document(**kwargs):
atomic_configs = AtomicConfig.objects(__raw__=kwargs).order_by(
'-_id').all()
for atomic_config in atomic_configs:
atomic_config.file_payload = fsc.retrieve(atomic_config.file_uid)
yield atomic_config
def run_simulation(sim):
# Load info from simulation request
sim_params, = find_simulation_parameter_document(_id=sim.params.id)
# TODO: Throw in some statments about timeouts, acceptable U(q), etc.
iterations = sim_params.iterations
target_acceptance = sim_params.target_acceptance
ensemble_temp = sim_params.temperature
# Load Starting Atoms
starting_atoms_entry, = find_atomic_config_document(
_id=sim.starting_atoms.id)
starting_atoms = starting_atoms_entry.file_payload
try:
traj_entry, = find_atomic_config_document(_id=sim.starting_atoms.id)
traj = traj_entry.file_payload
except:
traj_entry = None
traj = None
# We want to continue this simulation
if isinstance(traj, list):
# Give back the final configuration
atoms = traj[-1]
# Search filestore and get the file_location
with handler_context({'ase': FileLocation}):
atoms_file_location = fsc.retrieve(starting_atoms_entry.file_uid)
wtraj = PickleTrajectory(atoms_file_location, 'a')
# This is a new sim with a new trajectory
elif traj is None:
# Give back the initial config
atoms = starting_atoms
# Generate new file location and save it to filestore
new_atoms_entry = insert_atom_document(
starting_atoms_entry.name + '_' + sim.name, atoms)
with handler_context({'ase': FileLocation}):
new_file_location = fsc.retrieve(new_atoms_entry.file_uid)
wtraj = PickleTrajectory(new_file_location, 'w')
sim.simulation_atoms = new_atoms_entry
sim.save()
# Create Calculators
pes, = find_pes_document(_id=sim.pes.id)
master_calc = pes.payload
# Attach MulitCalc to atoms
atoms.set_calculator(master_calc)
sim.start_total_energy.append(atoms.get_total_energy())
sim.start_potential_energy.append(atoms.get_potential_energy())
sim.start_kinetic_energy.append(atoms.get_kinetic_energy())
sim.start_time.append(ttime.time())
sim.ran = True
sim.save()
# Simulate
# TODO: eventually support different simulation engines
out_traj, samples, l_p_i, seed = nuts(atoms, target_acceptance, iterations,
ensemble_temp, wtraj)
sim.end_time.append(ttime.time())
sim.total_iterations.append(sim.params.iterations)
sim.total_samples.append(samples)
sim.leapfrog_per_iter.append(l_p_i)
sim.finished = True
sim.seed.append(seed)
sim.save()
# Write info to DB
sim.final_potential_energy.append(out_traj[-1].get_potential_energy())
sim.final_kinetic_energy.append(out_traj[-1].get_kinetic_energy())
sim.final_kinetic_energy.append(out_traj[-1].get_total_energy())
sim.save()
def _npsave_helper(dd, base_path):
eid = fs_write.save_ndarray(dd, base_path)
with fsr.handler_context({'npy': fs_read.NpyHandler}):
ret = fsc.retrieve(eid)
assert_array_equal(dd, ret)
def run_simulation(sim):
# Load info from simulation request
sim_params, = find_simulation_parameter_document(_id=sim.params.id)
# TODO: Throw in some statments about timeouts, acceptable U(q), etc.
iterations = sim_params.iterations
target_acceptance = sim_params.target_acceptance
ensemble_temp = sim_params.temperature
# Load Starting Atoms
starting_atoms_entry, = find_atomic_config_document(_id=sim.starting_atoms.id)
starting_atoms = starting_atoms_entry.file_payload
try:
traj_entry, = find_atomic_config_document(_id=sim.starting_atoms.id)
traj = traj_entry.file_payload
except:
traj_entry = None
traj = None
# We want to continue this simulation
if isinstance(traj, list):
# Give back the final configuration
atoms = traj[-1]
# Search filestore and get the file_location
with handler_context({'ase': FileLocation}):
atoms_file_location = fsc.retrieve(starting_atoms_entry.file_uid)
wtraj = PickleTrajectory(atoms_file_location, 'a')
# This is a new sim with a new trajectory
elif traj is None:
# Give back the initial config
atoms = starting_atoms
# Generate new file location and save it to filestore
new_atoms_entry = insert_atom_document(
starting_atoms_entry.name + '_' + sim.name, atoms)
with handler_context({'ase': FileLocation}):
new_file_location = fsc.retrieve(new_atoms_entry.file_uid)
wtraj = PickleTrajectory(new_file_location, 'w')
sim.simulation_atoms = new_atoms_entry
sim.save()
# Create Calculators
pes, = find_pes_document(_id=sim.pes.id)
master_calc = pes.payload
# Attach MulitCalc to atoms
atoms.set_calculator(master_calc)
sim.start_total_energy.append(atoms.get_total_energy())
sim.start_potential_energy.append(atoms.get_potential_energy())
sim.start_kinetic_energy.append(atoms.get_kinetic_energy())
sim.start_time.append(ttime.time())
sim.ran = True
sim.save()
# Simulate
# TODO: eventually support different simulation engines
out_traj, samples, l_p_i, seed = nuts(atoms, target_acceptance, iterations,
ensemble_temp, wtraj)
sim.end_time.append(ttime.time())
sim.total_iterations.append(sim.params.iterations)
sim.total_samples.append(samples)
sim.leapfrog_per_iter.append(l_p_i)
sim.finished = True
sim.seed.append(seed)
sim.save()
# Write info to DB
sim.final_potential_energy.append(out_traj[-1].get_potential_energy())
sim.final_kinetic_energy.append(out_traj[-1].get_kinetic_energy())
sim.final_kinetic_energy.append(out_traj[-1].get_total_energy())
sim.save()
