def hilighter(event):
# if we did not hit a node, bail
if not hasattr(event, 'nodes') or not event.nodes:
return
# pull out the graph,
graph = event.artist.graph
# clear any non-default color on nodes
for node, attributes in graph.nodes.data():
attributes.pop('color', None)
for u, v, attributes in graph.edges.data():
attributes.pop('width', None)
for node in event.nodes:
graph.nodes[node]['color'] = 'C1'
#print(node)
for edge_attribute in graph[node].values():
edge_attribute['width'] = 3
# update the screen
event.artist.stale = True
event.artist.figure.canvas.draw_idle()
print(event.nodes)
simulation_id = event.nodes[0]
simulation = simulations[str(simulation_id)]
pp(simulation)
print("simulation id", simulation_id)
atoms = atoms_dict_to_ase(simulation["atoms"])
view(atoms)
def run_task(self, fw_spec):
ase_atoms_lst = self["ase_atoms_lst"]
workflow_id = fw_spec.get("workflow", {"_id": -1}).get("_id", -1)
logging.debug(fw_spec)
update_spec = fw_spec
calc_ids = []
for atoms_dict in ase_atoms_lst:
atoms = atoms_dict_to_ase(atoms_dict)
dct = atoms.info
total_energy = dct.get(
"E",
dct.get(
"energy",
dct.get(
"total_energy",
dct.get("TotalEnergy", dct.get("totalenergy",
"UNKNOWN")))))
if total_energy == "UNKNOWN":
logging.warning(
"total energy of cluster not specified in structure file")
nanocluster_atom_ids = list(range(len(atoms)))
db = get_external_database(fw_spec["extdb_connect"])
simulations = db['simulations']
# request id counter
simulation_id = _query_id_counter_and_increment('simulations', db)
nanocluster = {
"atom_ids": nanocluster_atom_ids,
"reference_id": simulation_id
}
dct = {
"_id": simulation_id,
"atoms": atoms_dict,
"source_id": -1,
"workflow_id": workflow_id,
"nanoclusters": [nanocluster],
"adsorbates": [],
"substrates": [],
"operations": [],
"inp": {},
"output": {
"total_energy": total_energy
},
}
# update external database
simulations.insert_one(dct)
calc_ids.append(simulation_id)
# update temp workflow data
update_spec["temp"]["calc_ids"] = calc_ids
update_spec.pop("_category")
update_spec.pop("name")
return FWAction(update_spec=update_spec)
def onpick3(event):
ind = event.ind
print('onpick3 scatter:', ind)
try:
first_id = ind[0]
except:
return
print("taking first point", first_id)
simulation = initial_guess_simulations_list[first_id]
pp(simulation)
print("simulation id", simulation["_id"])
atoms = atoms_dict_to_ase(simulation["atoms"])
view(atoms)
def onpick3(event):
ind = event.ind
print('onpick3 scatter:', ind)
try:
first_id = ind[0]
except:
return
#print(first_id)
simulation_id = dft_ids[first_id]
simulation = simulations[str(simulation_id)]
pp(simulation)
print("simulation id", simulation_id)
atoms = atoms_dict_to_ase(simulation["atoms"])
view(atoms)
def gather_all_atom_types(calc_ids, simulations):
"""Helper function to determine all atom types in the dataset
Args:
calc_ids (list) : ids of the simulation collection
simulations (list) : simulation documents
Returns:
list : a sorted unique list of atomic numbers in the
dataset
"""
# going through nc atoms once to find atom types
atomic_numbers = []
for idx, calc_id in enumerate(calc_ids):
atoms_dict = simulations[str(calc_id)]["atoms"]
atoms = atoms_dict_to_ase(atoms_dict)
atomic_numbers.extend(atoms.get_atomic_numbers())
sorted_list_atomic_numbers = list(sorted(set(atomic_numbers)))
all_atomtypes = sorted_list_atomic_numbers
return all_atomtypes
i = 0
epoch = 0
while i < MAX_VERTICAL:
current_simulations = odict()
print("# simulations on current level", len(children_ids))
print(children_ids)
for idx in children_ids:
current_simulations[idx] = simulations[idx]
current_structures = odict()
for idx, simulation in current_simulations.items():
step = simulation["operations"][0]
#print("step:", step)
atoms = atoms_dict_to_ase(simulation["atoms"])
current_structures[idx] = atoms
n_adsorbates = len(atoms[atoms.get_atomic_numbers() != 0]) - 55
#print(n_adsorbates)
# compare parent with current structures
marked_structures = odict()
if not is_root: # add safeguard for root and reprints
for idx, atoms in current_structures.items():
# mapping of parents to children
parent_idx = current_simulations[idx]["source_id"]
step = current_simulations[idx]["operations"][0]
parent_atoms = parent_structures[str(parent_idx)]
# shortened list of adsorption energies (possibly the workflow contains more than two different elements)
y_property_shortened.append(property_dct[latest_simulation["_id"]])
compositions = np.array(compositions)
# summarized structures of nanoclusters fully covered with adsorbate #
######################################################################
covered_structures = []
nanoclusters = []
# sort nanoclusters by element 1
nel1_lst = []
for nanocluster in ref_simulations_list:
nano_atoms_dct = nanocluster["atoms"]
nano_atoms = atoms_dict_to_ase(nano_atoms_dct)
nel1 = (nano_atoms.get_atomic_numbers() == element1).sum()
nel1_lst.append(nel1)
sort_ids = np.argsort(np.array(nel1_lst))
ref_simulations_list = np.array(ref_simulations_list)[sort_ids]
#print("ref_simulations_list", ref_simulations_list)
for nanocluster in ref_simulations_list:
ref_id = nanocluster["_id"]
nano_atoms_dct = nanocluster["atoms"]
nano_atoms = atoms_dict_to_ase(nano_atoms_dct)
covered_atoms = nano_atoms.copy()
related_ids = reference_id_dct[ref_id]
for related_id in related_ids:
positions = positions_dct[related_id]