def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
fixed_points = None
if self.simulation.fixpoints is not None:
fixed_points = loads(self.simulation.fixpoints)
# print(fixed_points)
nodes = list(fixed_points.keys())[3:]
new_fps = []
for i in range(len(fixed_points['FP'])):
new_fps.append({
"proba": fixed_points['Proba'][str(i)],
"nodes":
{node: fixed_points[node][str(i)]
for node in nodes}
})
fixed_points = new_fps
return Response(
{
'fixed_points': fixed_points,
'status': self.simulation.status
},
status=status.HTTP_200_OK)
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
result = {
'done': self.simulation.status == MaBoSSSimulation.ENDED,
'failed': self.simulation.status in [MaBoSSSimulation.ERROR, MaBoSSSimulation.INTERRUPTED]
}
if self.simulation.status == MaBoSSSimulation.ERROR:
result.update({'error': self.simulation.error})
return Response(result, status=status.HTTP_200_OK)
def get(self, request, project_id, simulation_id): HasMaBoSSSimulation.load(self, request, project_id, simulation_id) cfg_filename = self.getCFGFilePath() if cfg_filename is None: return Response(status=status.HTTP_501_NOT_IMPLEMENTED) return FileResponse( open(cfg_filename, 'rb'), as_attachment=True, filename=basename(cfg_filename) )
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
if self.simulation.nodes_probtraj is not None:
nodes_probtraj = loads(self.simulation.nodes_probtraj)
else:
nodes_probtraj = None
return Response(
{
'nodes_probtraj': nodes_probtraj,
'status': self.simulation.status
},
status=status.HTTP_200_OK)
def get(self, request, project_id, simulation_id): HasMaBoSSSimulation.load(self, request, project_id, simulation_id) try: new_model = LogicalModel( project=self.project, name=self.simulation.name + " model", bnd_file=File(open(self.getBNDFilePath(), 'rb'), name=basename(self.getBNDFilePath())), cfg_file=File(open(self.getCFGFilePath(), 'rb'), name=basename(self.getCFGFilePath())), format=LogicalModel.MABOSS ) new_model.save() return Response(status=status.HTTP_200_OK) except: return Response(status=status.HTTP_500_INTERNAL_SERVER_ERROR)
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
fixed_points = None
if self.simulation.states_probtraj is not None:
fixed_points = {}
for key, values in loads(self.simulation.states_probtraj).items():
last_time = list(values.keys())[len(values.keys()) - 1]
if values[last_time] > 0.01:
fixed_points.update({key: values[last_time]})
return Response(
{
'last_states': fixed_points,
'status': self.simulation.status
},
status=status.HTTP_200_OK)
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
if self.simulation.states_probtraj is not None:
states_probtraj = {}
for key, value in loads(self.simulation.states_probtraj).items():
values = list(value.values())
if max(values) > 0.01:
states_probtraj.update({key: value})
# states_probtraj = {key: value for key, value in loads(self.simulation.states_probtraj).items() if list(value.values())[len(value.values()) - 1] > 0.01}
else:
states_probtraj = None
return Response(
{
'states_probtraj': states_probtraj,
'status': self.simulation.status
},
status=status.HTTP_200_OK)
def delete(self, request, project_id, simulation_id): HasMaBoSSSimulation.load(self, request, project_id, simulation_id) self.simulation.delete() return Response(status=status.HTTP_200_OK)
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
model = maboss.load(self.getBNDFilePath(), self.getCFGFilePath())
if self.simulation.fixpoints is not None:
fixed_points = loads(self.simulation.fixpoints)
output_vars = [
label for label, node in model.network.items()
if not node.is_internal
]
nb_nodes = len(output_vars)
res = np.zeros((nb_nodes, 0))
for i, proba in enumerate(fixed_points["Proba"]):
t_res = np.zeros((nb_nodes, 1))
for node in output_vars:
t_res[output_vars.index(node)] = fixed_points[node][str(i)]
res = np.append(res, t_res, axis=1)
mat = np.transpose(res)
pca = PCA()
pca_res = pca.fit(mat)
x_new = pca.transform(mat)
res = {}
for i, state in enumerate(fixed_points["State"]):
res.update({fixed_points["State"][state]: list(x_new[i, 0:2])})
raw_arrows = np.transpose(pca_res.components_[0:2, :])
# Adjusting arrow length to fit in the graph
max_x_arrows = max(raw_arrows[:, 0])
min_x_arrows = min(raw_arrows[:, 0])
max_x_values = max(x_new[:, 0])
min_x_values = min(x_new[:, 0])
min_x_ratio = min_x_values / min_x_arrows
max_x_ratio = max_x_values / max_x_arrows
x_ratio = min(min_x_ratio, max_x_ratio)
max_y_arrows = max(raw_arrows[:, 1])
min_y_arrows = min(raw_arrows[:, 1])
max_y_values = max(x_new[:, 1])
min_y_values = min(x_new[:, 1])
min_y_ratio = min_y_values / min_y_arrows
max_y_ratio = max_y_values / max_y_arrows
y_ratio = min(min_y_ratio, max_y_ratio)
from math import log10, floor
def round_sig(x, sig=2):
if x == 0.0:
return x
return round(x, sig - int(floor(log10(abs(x)))) - 1)
values = []
names = []
for i, arrow_raw in enumerate(raw_arrows):
as_list = [round(val, 2) for val in list(arrow_raw)]
if as_list not in values:
values.append(as_list)
names.append([output_vars[i]])
else:
names[values.index(as_list)].append(output_vars[i])
values = [[value[0] * x_ratio, value[1] * y_ratio]
for value in values]
return Response({
'status':
'Finished',
'data':
json.dumps(res),
'arrows':
values,
'arrowlabels':
names,
'explainedVariance': [
round_sig(value * 100, 4)
for value in list(pca.explained_variance_ratio_[0:2])
]
})
else:
return Response({'status': 'Busy'})
def get(self, request, project_id, simulation_id):
HasMaBoSSSimulation.load(self, request, project_id, simulation_id)
model = maboss.load(self.getBNDFilePath(), self.getCFGFilePath())
if self.simulation.states_probtraj is not None:
states_probtraj = loads(self.simulation.states_probtraj)
states_lastprob = {}
for state, values in states_probtraj.items():
list_values = list(values.values())
states_lastprob.update(
{state: list_values[len(list_values) - 1]})
nodes = [
name for name, node in model.network.items()
if not node.is_internal
]
nb_nodes = len(nodes)
nb_fp = len(states_lastprob.keys())
states = []
res = np.zeros((nb_nodes, 0))
for index, proba in states_lastprob.items():
if proba > 0.01:
states.append(index)
if index == "<nil>":
res = np.append(res, np.zeros((nb_nodes, 1)), axis=1)
else:
t_res = np.zeros((nb_nodes, 1))
for node in [
t_node.strip() for t_node in index.split("--")
]:
# print(node)
t_res[nodes.index(node)] = 1
res = np.append(res, t_res, axis=1)
columns = ["FP%d" % i for i in range(res.shape[1])]
mat = np.transpose(res)
pca = PCA()
pca_res = pca.fit(mat)
x_new = pca.transform(mat)
res = {}
for i, state in enumerate(states):
res.update({state: list(x_new[i, 0:2])})
raw_arrows = np.transpose(pca_res.components_[0:2, :])
# Adjusting arrow length to fit in the graph
max_x_arrows = max(raw_arrows[:, 0])
min_x_arrows = min(raw_arrows[:, 0])
max_x_values = max(x_new[:, 0])
min_x_values = min(x_new[:, 0])
min_x_ratio = abs(min_x_values / min_x_arrows)
max_x_ratio = abs(max_x_values / max_x_arrows)
x_ratio = min(min_x_ratio, max_x_ratio)
max_y_arrows = max(raw_arrows[:, 1])
min_y_arrows = min(raw_arrows[:, 1])
max_y_values = max(x_new[:, 1])
min_y_values = min(x_new[:, 1])
min_y_ratio = abs(min_y_values / min_y_arrows)
max_y_ratio = abs(max_y_values / max_y_arrows)
y_ratio = min(min_y_ratio, max_y_ratio)
from math import log10, floor
def round_sig(x, sig=2):
if x == 0.0:
return x
return round(x, sig - int(floor(log10(abs(x)))) - 1)
values = []
names = []
for i, arrow_raw in enumerate(raw_arrows):
as_list = [round(val, 2) for val in list(arrow_raw)]
if as_list not in values:
values.append(as_list)
names.append([nodes[i]])
else:
names[values.index(as_list)].append(nodes[i])
values = [[value[0] * x_ratio, value[1] * y_ratio]
for value in values]
return Response({
'status':
'Finished',
'data':
json.dumps(res),
'arrows':
values,
'arrowlabels':
names,
'explainedVariance': [
round_sig(value * 100, 4)
for value in list(pca.explained_variance_ratio_[0:2])
]
})
else:
return Response({'status': 'Busy'})
