def test_game6(self): game6 = BayesianModel([ ("X", "F"), ("C", "X"), ("A", "C"), ("A", "D"), ("B", "D"), ("B", "E"), ("D", "X"), ("D", "Y"), ("E", "Y"), ("F", "Y"), ]) inference = CausalInference(game6) self.assertFalse(inference.is_valid_backdoor_adjustment_set("X", "Y")) deconfounders = inference.get_all_backdoor_adjustment_sets("X", "Y") self.assertEqual( deconfounders, frozenset({ frozenset({"C", "D"}), frozenset({"A", "D"}), frozenset({"D", "E"}), frozenset({"B", "D"}), }), )
def test_game3(self): game3 = BayesianModel([("X", "Y"), ("X", "A"), ("B", "A"), ("B", "Y"), ("B", "X")]) inference = CausalInference(game3) self.assertFalse(inference.is_valid_backdoor_adjustment_set("X", "Y")) deconfounders = inference.get_all_backdoor_adjustment_sets("X", "Y") self.assertEqual(deconfounders, frozenset({frozenset({"B"})}))
def test_create_estimator(self): game1 = BayesianModel([("X", "A"), ("A", "Y"), ("A", "B")]) data = pd.DataFrame(np.random.randint(2, size=(1000, 4)), columns=["X", "A", "B", "Y"]) inference = CausalInference(model=game1) ate = inference.estimate_ate("X", "Y", data=data, estimator_type="linear") self.assertAlmostEqual(ate, 0, places=0)
def sample_dag(dag, num): #zzz this loses disconnected nodes!!! # bayesmod = BayesianModel(dag.edges()) # bayesmod = BayesianModel(dag) bayesmod = BayesianModel() bayesmod.add_nodes_from(dag.nodes()) bayesmod.add_edges_from(dag.edges()) tab_cpds = [] cards = {node: len(dag.node[node]['cpd']) for node in dag.nodes()} for node in dag.nodes(): parents = dag.predecessors(node) cpd = dag.node[node]['cpd'] if parents: parent_cards = [cards[par] for par in parents] logging.debug("TablularCPD({}, {}, {}, {}, {})".format( node, cards[node], cpd, parents, parent_cards)) tab_cpds.append( TabularCPD(node, cards[node], cpd, parents, parent_cards)) else: logging.debug("TablularCPD({}, {}, {})".format( node, cards[node], cpd)) tab_cpds.append(TabularCPD(node, cards[node], cpd)) logging.debug("cpds add: {}".format(tab_cpds)) print "model variables:", bayesmod.nodes() for tab_cpd in tab_cpds: print "cpd variables:", tab_cpd.variables bayesmod.add_cpds(*tab_cpds) logging.debug("cpds get: {}".format(bayesmod.get_cpds())) inference = BayesianModelSampling(bayesmod) logging.debug("generating data") recs = inference.forward_sample(size=num, return_type='recarray') return recs
def test_game2(self): game2 = BayesianModel([ ("X", "E"), ("E", "Y"), ("A", "B"), ("A", "X"), ("B", "C"), ("D", "B"), ("D", "E"), ]) inference = CausalInference(game2) self.assertTrue(inference.is_valid_backdoor_adjustment_set("X", "Y")) deconfounders = inference.get_all_backdoor_adjustment_sets("X", "Y") self.assertEqual(deconfounders, frozenset())
from pgmpy.models.BayesianModel import BayesianModel from pgmpy.factors.discrete import TabularCPD from pgmpy.sampling import BayesianModelSampling student = BayesianModel([('diff', 'grade'), ('intel', 'grade')]) cpd_d = TabularCPD('diff', 2, [[0.6], [0.4]]) cpd_i = TabularCPD('intel', 2, [[0.7], [0.3]]) cpd_g = TabularCPD('grade', 3, [[0.3, 0.05, 0.9, 0.5], [0.4, 0.25,0.08, 0.3], [0.3, 0.7, 0.02, 0.2]],['intel', 'diff'], [2, 2]) student.add_cpds(cpd_d, cpd_i, cpd_g) inference = BayesianModelSampling(student) print inference.forward_sample(size=3, return_type='recarray')
entropy = {} for i in range(df.shape[1]): entropy[list( df[[i]])[0]] = -1 * np.sum(norm.logpdf(df[[i]])) / df.shape[0] mean[list(df[[i]])[0]] = np.mean(df[[i]]) print "Training Neural Network for continous variables" #uncomment this nnw = neuralNet.trainNeuralNetwork() file1 = open("cmd.txt") file2 = file.read(file1) #model = eval(file2) print "Creating Bayesian Model" model = BayesianModel(getLinks.getLinksOfNodes()) #print("Making parents dictionary") parents = getLinks.getParents() pe = ParameterEstimator(model, df) nodes = getNodes.getNodesFromCSV() cpds = {} variableCard = {} for node in nodes: try: cpd = pe.state_counts(node) cpd = cpd.transpose() cpd_prob = cpd.div(cpd.sum(axis=1), axis=0)
def setUp(self): self.game = BayesianModel([("A", "X"), ("A", "B"), ("C", "B"), ("C", "Y"), ("X", "Y"), ("B", "X")]) self.inference = CausalInference(self.game)
from pgmpy.factors.discrete import State from pgmpy.models.BayesianModel import BayesianModel from pgmpy.factors.discrete import TabularCPD from pgmpy.sampling import BayesianModelSampling from pgmpy.inference import VariableElimination burglary = BayesianModel([('Burglary', 'Alarm'), ('Earthquake', 'Alarm'), ('Alarm', 'Johncalls'), ('Alarm', 'Marycalls')]) cpd_burgary = TabularCPD('Burglary', 2, [[0.001], [0.999]]) cpd_earthquake = TabularCPD('Earthquake', 2, [[0.002], [0.998]]) cpd_alarm = TabularCPD('Alarm', 2, [[0.99, 0.71, 0.06, 0.05], [0.01, 0.29, 0.94, 0.95]], ['Burglary', 'Earthquake'], [2, 2]) cpd_johncall = TabularCPD('Johncalls', 2, [[0.95, 0.1], [0.05, 0.90]], ['Alarm'], [2]) cpd_marrycall = TabularCPD('Marycalls', 2, [[0.99, 0.30], [0.01, 0.70]], ['Alarm'], [2]) burglary.add_cpds(cpd_burgary, cpd_earthquake, cpd_alarm, cpd_johncall, cpd_marrycall) for cpd in burglary.get_cpds(): print(cpd) ####################################################################### """inference using variable elimination""" inference = VariableElimination(burglary) alarm = inference.query(['Alarm'], {'Johncalls': 0}) print(alarm['Alarm']) ##############################################################
class Graph(object): ## CLASS TO GENERATE The Dynamic bayes network and to find the probability of success of an plan. def __init__(self): ####constructor method initialize the object self.plans=[] self.final=[] self.final1=[] self.final2=[] self.nodes_len=0 self.my_dic={} self.my_dic1={} self.my_dic2={} self.received=False self.grounded=[] self.action_values={} self.actions2=0 self.statenodes=[] def total_plan(self,data): ##### calll back for the the total order plans from oscar. msg=data total_plans=len(msg.esterel_plans) self.nodes_len=len(msg.esterel_plans[0].nodes) print(total_plans) rospy.loginfo("total plan received :" + str(total_plans)) rospy.loginfo("probability :" + str(len(msg.plan_success_prob))) for i in range(len(msg.esterel_plans)):###for all the total order plans new=[] for j in range(len(msg.esterel_plans[i].nodes)):#### Actions in each plan new.append(msg.esterel_plans[i].nodes[j].name) self.grounded.append(msg.esterel_plans[i].nodes[j].action.name) # val.append(msg.esterel_plans[i].nodes[j].action.parameters[i].) ### each plan with actions array of array. self.my_dic1[i]=(msg.esterel_plans[i].nodes[j].name +" "+ str(msg.esterel_plans[i].nodes[j].action.parameters)) self.plans.append(new) size= len(msg.esterel_plans[0].nodes[0].action.parameters) print(size) ########### Total order plans actions with parameters ########################################## for i in range(len(msg.esterel_plans)): for j in range(len(msg.esterel_plans[i].nodes)): self.my_dic[j]= msg.esterel_plans[i].nodes[j].name for k in range(len(msg.esterel_plans[i].nodes[j].action.parameters)): self.my_dic[j] += " " + msg.esterel_plans[i].nodes[j].action.parameters[k].value # print(self.my_dic[j]) for i in range(len(self.my_dic)): # if i % 2==0: self.final.append(self.my_dic[i]) self.final1.append(self.final[i].split()) self.final2.append('#'.join(self.final1[i])+'%'+str(i)) # list1=plan+val1 self.received=True # self.final.append(str(' '.join(plan))) # print("totalplans",self.action_values) print("totalplans",self.final2) # print("grounded actions",self.grounded) # print("totalplans",self.my_dic) # print("dictionary with parameters of total plans",self.my_dic) # print("PLAN",plan) # print("values",val1) def dump(self): f= open("predicates.txt","w+") print("avinash") f.write("state nodes for each layer \n") for i in range(len(self.statenodes)): f.write("Layer No :"+ str(i)+"\n") for j in range(len(self.statenodes[i])): f.write( self.statenodes[i][j]+"\n") f.close() def parents(self): f= open("parents_and_children.txt","w+") f.write("Parents nodes for each layer \n") for i in range(len(self.statenodes)): for j in range(len(self.statenodes[i])): f.write("Node_name :" +self.statenodes[i][j] +"\n") parent=self.model.get_parents(self.statenodes[i][j]) f.write("Parent:"+ str(parent)+ "\n") f.write("\n") for i in range(self.actions2): f.write("Parents for actions \n") f.write("Node_name :" +self.final2[i] +"\n") parent=self.model.get_parents(self.final2[i]) f.write("Parent:"+ str(parent)+ "\n") f.write("\n") for i in range(self.actions2): f.write("Children for actions \n") f.write("Node_name :" +self.final2[i]+"\n") child=self.model.get_children(self.final2[i]) f.write("Children:"+ str(child)+ "\n") f.write("\n") def call_service(self): #### subscribing to the node to get the total order plans.#################################### rospy.Subscriber("/csp_exec_generator/valid_plans", EsterelPlanArray, self.total_plan) print ("Waiting for service") rospy.wait_for_service('/rosplan_knowledge_base/domain/name') rospy.wait_for_service('/rosplan_knowledge_base/domain/types') # rospy.wait_for_service('/rosplan_knowledge_base/domain/functions') rospy.wait_for_service('/rosplan_knowledge_base/domain/operators') rospy.wait_for_service('/rosplan_knowledge_base/domain/operator_details') # rospy.wait_for_service('/rosplan_knowledge_base/domain/predicate_details') rospy.wait_for_service('/rosplan_knowledge_base/domain/predicates') rospy.wait_for_service('/rosplan_knowledge_base/state/propositions') # rospy.wait_for_service('/rosplan_knowledge_base/state/timed_knowledge') # rospy.wait_for_service('/rosplan_knowledge_base/query_state') # # try: print ("Calling Service") domain_name = rospy.ServiceProxy('/rosplan_knowledge_base/domain/name', GetDomainNameService) domain_types = rospy.ServiceProxy('/rosplan_knowledge_base/domain/types',GetDomainTypeService) # domain_functions = rospy.ServiceProxy('/rosplan_knowledge_base/domain/functions', GetDomainAttributeService) domain_operatordetails = rospy.ServiceProxy('/rosplan_knowledge_base/domain/operator_details', GetDomainOperatorDetailsService) domain_operators = rospy.ServiceProxy('/rosplan_knowledge_base/domain/operators', GetDomainOperatorService) domain_predicates = rospy.ServiceProxy('/rosplan_knowledge_base/domain/predicates', GetDomainAttributeService) problem_initialstate = rospy.ServiceProxy('/rosplan_knowledge_base/state/propositions',GetAttributeService) # domain_timedknowledge = rospy.ServiceProxy('/rosplan_knowledge_base/state/timed_knowledge',GetAttributeService) # query_proxy = rospy.ServiceProxy('/rosplan_knowledge_base/query_state', KnowledgeQueryService) resp1 = domain_name() resp2 = domain_types() resp5 = domain_operators() # resp3 = domain_functions() # resp4.op.formula.typed_parameters[0].key # resp4 = domain_operatordetails(act) resp7 = domain_predicates() resp8 = problem_initialstate() # print("resp7", resp7) # print ("action_names",resp4.op.formula.name) #actions act_nodes=[] actions=len(resp5.operators) print("actions",actions) for i in range(actions): act_nodes.append(str(resp5.operators[i].name)) print("action_nodes",act_nodes) print("waiting for the message") while self.received is False : continue TotalPlans=self.plans # finalplans=self.final finalplans1=self.final2 # # print(TotalPlans[0][:]) ncol=len(TotalPlans[0]) # print("coloums",ncol) durative=[] Actions_wth_param=[] for i in range(1): for j in range(ncol): Actions_wth_param.append(TotalPlans[i][j]) if j % 2 == 0: durative.append(TotalPlans[i][j]) actions1=len(durative)#### Not every action 17 self.actions2=len(Actions_wth_param)### every action 34 print("durative actions",actions1) print("durative actions",durative) # print("actions_plan",Actions_wth_param) grounded=[] grounded1=[] for i in range(len(self.grounded)): grounded1.append(self.grounded[i]) if i%2!=0: grounded.append(self.grounded[i]) # print("grounded",grounded) # print("grounded",grounded1) ######## start actions and end actions from a plan start_actions=[] end_actions=[] for i in range(len(self.final)): if i%2==0: start_actions.append(self.final[i]) else: end_actions.append(self.final[i]) start_actions1=[] end_actions1=[] for i in range(len(self.final2)): if i%2==0: start_actions1.append(self.final2[i]) else: end_actions1.append(self.final2[i]) # print("start_actions",start_actions) # print("end_actions",end_actions ) start_cond=[] end_cond=[] overall_cond=[] at_start_eff=[] at_start_deleff=[] at_end_addeff=[] at_end_deleff=[] operator_details=[] operator_detail=[] values=[] names=[] start=[] end=[] overall=[] start_param=[] end_param=[] overall_param=[] start_eff=[] start_del=[] end_eff=[] end_del=[] add_param=[] del_param=[] end_del_param=[] variable=[] grounded_parameteres=[] start1=[] res=[] # start_strings =[] end_strings =[] overall_strings =[] res2 =[] str1=[] str2=[] str3=[] str4=[] str5=[] str6=[] str7=[] all_set=set() strt_c={} self.model=BayesianModel() self.dbn=DBN() ################### start actions ############################################### for i in range(self.actions2): ######## action nodes from total order plan ############################## self.model.add_node(finalplans1[i]+'%'+str(i)) ################### ACTION PREDICATES FIELD######################################################## for i in range(actions1): start_strings =[] end_strings=[] overall_strings=[] start_param=[] end_param=[] add_param=[] end_del_param=[] resp4 =domain_operatordetails(grounded[i]) operator_details.append(str(resp4.op.formula.name)) # model.add_node(finalplans[i]) #################### Parameters of the action ########################################################## size=len(resp4.op.formula.typed_parameters) # print(size) for j in range(size): operator_details[i] =operator_details[i]+ " "+ resp4.op.formula.typed_parameters[j].key detail = operator_details[i].split() action_name = detail[0] var_Dic = dict(zip(detail,start_actions[i].split())) del var_Dic[action_name] # var_Dic=str(var_Dic) # print("dictionary",str(var_Dic)) variable.append(var_Dic) #####Dictionary getting the variables for replacement in the preconditions and effects ########## at_start conditions################################################################### size=len(resp4.op.at_start_simple_condition) # print("preconditions length",size) for z in range(size): start=[] if size!=0 : start.append(str(resp4.op.at_start_simple_condition[z].name)) start1.append(str(resp4.op.at_start_simple_condition[z].name)) for j in range(len(resp4.op.at_start_simple_condition[z].typed_parameters)): start.append(str(resp4.op.at_start_simple_condition[z].typed_parameters[j].key)) start1.append(str(resp4.op.at_start_simple_condition[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in start] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # print(res_index) # res = start for idx,val in enumerate(res_index): start[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str1.append('#'.join(start)) #+'%'+str(0)) start_strings.append('#'.join(start))#+'%'+str(0)) start_cond.append(start_strings) ########## at_end conditions##################################################################### size=len(resp4.op.at_end_simple_condition) # print("nnnnn",size) for z in range(size): end=[] if size!=0: end.append(str(resp4.op.at_end_simple_condition[z].name)) for j in range(len(resp4.op.at_end_simple_condition[z].typed_parameters)): end.append(str(resp4.op.at_end_simple_condition[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in end] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) # res1 = end for idx,val in enumerate(res_index): end[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str2.append('#'.join(end))# +'%'+str(0)) end_strings.append('#'.join(end))# +'%'+str(i)) end_cond.append(end_strings) ########## overall conditions#################################################################### size=len(resp4.op.over_all_simple_condition) # print("nnnnn",size) for z in range(size): overall=[] if size!=0: overall.append(str(resp4.op.over_all_simple_condition[z].name)) for j in range(len(resp4.op.over_all_simple_condition[z].typed_parameters)): overall.append(str(resp4.op.over_all_simple_condition[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in overall] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) for idx,val in enumerate(res_index): overall[val] = var_update[idx] #### replacement done variables with parameters str3.append('#'.join(overall))#+'%'+str(0)) overall_strings.append('#'.join(overall))#+'%'+str(i)) overall_cond.append(overall_strings) ########## at_start_add_effects ################################################################### size=len(resp4.op.at_start_add_effects) # print("nnnnn",size) for z in range(size): start_eff=[] if size!=0: start_eff.append(str(resp4.op.at_start_add_effects[z].name)) for j in range(len(resp4.op.at_start_add_effects[z].typed_parameters)): start_eff.append(str(resp4.op.at_start_add_effects[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in start_eff] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) # res1 = end for idx,val in enumerate(res_index): start_eff[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str4.append('#'.join(start_eff))# +'%'+str(0)) start_param.append('#'.join(start_eff)) at_start_eff.append(start_param) ########## at_start_DEL_effects ################################################################ size=len(resp4.op.at_start_del_effects) # print("at start del effects",size) for z in range(size): start_del=[] if size!=0: start_del.append(str(resp4.op.at_start_del_effects[z].name)) for j in range(len(resp4.op.at_start_del_effects[z].typed_parameters)): start_del.append(str(resp4.op.at_start_del_effects[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in start_del] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) # res1 = end for idx,val in enumerate(res_index): start_del[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str5.append('#'.join(start_del))# +'%'+str(0)) end_param.append('_'.join(start_del))#+'%'+str(i)) at_start_deleff.append(end_param) ########## at_end_add_effects ##################################################################### size=len(resp4.op.at_end_add_effects) # print("nnnnn",size) for z in range(size): end_eff=[] if size!=0: end_eff.append(str(resp4.op.at_end_add_effects[z].name)) for j in range(len(resp4.op.at_end_add_effects[z].typed_parameters)): end_eff.append(str(resp4.op.at_end_add_effects[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in end_eff] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) # res1 = end for idx,val in enumerate(res_index): end_eff[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str6.append('#'.join(end_eff)) #+'%'+str(0)) add_param.append('#'.join(end_eff))#+'%'+str(i)) at_end_addeff.append(add_param) ########## at_end_del_effects ####################################################################### size=len(resp4.op.at_end_del_effects) # print("nnnnn",size) for z in range(size): end_del=[] if size!=0: end_del.append(str(resp4.op.at_end_del_effects[z].name)) for j in range(len(resp4.op.at_end_del_effects[z].typed_parameters)): end_del.append(str(resp4.op.at_end_del_effects[z].typed_parameters[j].key)) update=[var_Dic.get(key) for key in end_del] var_update = [i for i in update if i is not None] # print("l_update",var_update) res_index = [i for i in range(len(update)) if update[i] != None] # other than Nones positions # print(res_index) # res1 = end for idx,val in enumerate(res_index): end_del[val] = var_update[idx] #### replacement done variables with parameters dictionary(var_dic). str7.append('#'.join(end_del))# +'%'+str(0)) end_del_param.append('#'.join(end_del))# +'%'+str(i)) at_end_deleff.append(end_del_param) # result=dict(zip(operator_details[0].split(),self.final[0].split())) # result=set(result) # print("result",variable)####### Dictionary of parameters for replacement. print("final",res) # print("start_conditions",start_strings) joinedlist=str1+str2+str3+str4+str5+str6+str7 # joinedlist=start_strings+end_strings+overall_strings+start_param+end_param+add_param+end_del_param joinedlist1=start_cond+end_cond+overall_cond+start_eff+end_eff+at_start_eff+at_start_deleff+at_end_addeff+at_end_deleff # print("all set union",joinedlist1) # print("all set ",len(joinedlist)) # print("new",new) # print("operator_details",resp4) # print("start_cond",start_cond) # print(values) # print("parameters",operator_details) print("end_conditions names",end) print("overall_conditions names",overall) # print("start_conditions parameters",start_param) # # print("end_conditions parameters",end_param) # # print("overall_conditions parameters",overall_param) # print("at_start_add_effects_names",start_eff) # # print("at_start_add_effects_parameters",add_param) # print("at_start_delete_effects_names",start_del) # # print("at_start_delete_effects_parameters",del_param) # print("at_end_add_effects_names",end_eff) # # print("at_end_add_effects",end_param) # print("at_end_del_effects_names",end_del) # print("at_end_del_effects",end_del_param) length= (len(resp8. attributes)) # print(length) initialNodes=[] # self.statenodes=[] strings=[] ############## Initial nodes ####################################################################################### for i in range(length): initialNodes.append(str(resp8. attributes[i].attribute_name)) # key.append(resp8. attributes[i].values[0].value) strings.append(resp8. attributes[i].attribute_name) value_size = len(resp8. attributes[i].values) for j in range(value_size): strings[i] = strings[i] + "# " + resp8.attributes[i].values[j].value ### all the initial predicates from problems # initial.append(initialNodes[i]+'#'+key[i]) # print("initialnodes",str(initialNodes)) # print("Initial_nodes",strings)#initial nodes from problem pddl ######### state nodes in each transition of actions################################################################# for i in range(self.actions2+1): states=[] for nodes in joinedlist: states.append(str(nodes)+'%'+str(i)) self.statenodes.append(states) Graph.dump(self) #i+=1 # print("state_nodes",self.statenodes)# state_nodes dbn ####################################################################################################################### for i in range(len(self.statenodes)): for j in range(len(self.statenodes[i])): self.model.add_node(self.statenodes[i][j])############ All predicates nodes from Domain file ######################## ######state transitions with the initial nodes########################################### for i in range(self.actions2): for j in range(len(self.statenodes[i])): self.model.add_edge(self.statenodes[i][j], self.statenodes[i+1][j]) ######################################################################################## ### drawing edges from predicates to actions....######################################## for i in range(actions1): for j in range(len(start_cond[i])): self.model.add_edge(start_cond[i][j]+'%' +str(i),start_actions1[i]) # # model.add_edge(start_cond[i][j],self.final[i]) for i in range(actions1): for j in range(len(end_cond[i])): self.model.add_edge(end_cond[i][j]+'%' +str(i),end_actions1[i]) for i in range(actions1): for j in range(len(overall_cond[i])): self.model.add_edge(overall_cond[i][j]+'%' +str(i+1),end_actions1[i]) for i in range(actions1): for j in range(len(at_start_eff[i])): self.model.add_edge(start_actions1[i],at_start_eff[i][j]+'%' +str(i+1)) for i in range(actions1): for j in range(len(at_start_deleff[i])): self.model.add_edge(start_actions1[i],at_start_deleff[i][j]+'%' +str(i+1)) for i in range(actions1): for j in range(len(at_end_addeff[i])): self.model.add_edge(end_actions1[i],at_end_addeff[i][j]+'%' +str(i+2)) for i in range(actions1): for j in range(len(at_end_deleff[i])): self.model.add_edge(end_actions1[i],at_end_deleff[i][j]+'%' +str(i+2)) #################################################################################################################### ###### DBN Graph ######################## Graph.parents(self) np.warnings.filterwarnings('ignore') warnings.simplefilter(action="ignore",category=RuntimeWarning) # nx.draw(self.model, with_labels=True) plt.xlim(-0.05, 1.05) plt.ylim(-0.05, 1.05) plt.axis('off') x=self.model.nodes() print(len(x)) # plt.show() rospy.spin()
import networkx as nx import matplotlib.pyplot as plt from pgmpy.models.BayesianModel import BayesianModel # construct the tree graph structure model = BayesianModel([('A', 'B'), ('A', 'C'), ('B', 'D'), ('B', 'E'), ('C', 'F')]) nx.draw_circular(model, with_labels=True, arrowsize=30, node_size=800, alpha=0.3, font_weight='bold') plt.savefig('fig1.png', bbox_inches='tight') from pgmpy.factors.discrete import TabularCPD # add CPD to each edge cpd_a = TabularCPD('A', 2, [[0.4], [0.6]]) cpd_b = TabularCPD('B', 3, [[0.6, 0.2], [0.3, 0.5], [0.1, 0.3]], evidence=['A'], evidence_card=[2]) cpd_c = TabularCPD('C', 2, [[0.3, 0.4], [0.7, 0.6]], evidence=['A'], evidence_card=[2]) cpd_d = TabularCPD('D', 3, [[0.5, 0.3, 0.1], [0.4, 0.4, 0.8], [0.1, 0.3, 0.1]], evidence=['B'], evidence_card=[3])