def encoder(domains_list):
apn_list = []
for dom in domains_list:
for a in dom.apn:
if a not in apn_list:
apn_list.append(a)
save_arrays.save(apn_list, "./files/apn.obj")
cit_list = []
for dom in domains_list:
for c in dom.cit:
if c not in cit_list:
cit_list.append(c)
save_arrays.save(cit_list, "./files/cit.obj")
loc_list = []
for dom in domains_list:
for l in dom.loc:
if l not in loc_list:
loc_list.append(l)
save_arrays.save(loc_list, "./files/loc.obj")
obj_list = []
for dom in domains_list:
for o in dom.obj:
if o not in obj_list:
obj_list.append(o)
save_arrays.save(obj_list, "./files/obj.obj")
tru_list = []
for dom in domains_list:
for t in dom.tru:
if t not in tru_list:
tru_list.append(t)
save_arrays.save(tru_list, "./files/tru.obj")
def encoder(domains_list):
apn_list = []
for dom in domains_list:
apn_list = apn_list + [element.upper() for element in dom.apn]
apn_list = list(set(apn_list))
save_arrays.save(apn_list, "apn.obj")
cit_list = []
for dom in domains_list:
cit_list = cit_list + [element.upper() for element in dom.cit]
cit_list = list(set(cit_list))
save_arrays.save(cit_list, "cit.obj")
loc_list = []
for dom in domains_list:
loc_list = loc_list + [element.upper() for element in dom.loc]
loc_list = list(set(loc_list))
save_arrays.save(loc_list, "loc.obj")
obj_list = []
for dom in domains_list:
obj_list = obj_list + [element.upper() for element in dom.obj]
obj_list = list(set(obj_list))
save_arrays.save(obj_list, "obj.obj")
tru_list = []
for dom in domains_list:
tru_list = tru_list + [element.upper() for element in dom.tru]
tru_list = list(set(tru_list))
save_arrays.save(tru_list, "tru.obj")
'action_type': train_y[0],
'type1': train_y[1],
'param1': train_y[2],
'type2': train_y[3],
'param2': train_y[4],
'type3': train_y[5],
'param3': train_y[6],
'type4': train_y[7],
'param4': train_y[8]
},
batch_size=128,
epochs=10,
verbose=2,
validation_split=0.1)
save_arrays.save(history, 'history')
net.save('model')
else:
name = sys.argv[1]
model = load_model(name)
testing = open("mini_test", "rb")
test = pickle.load(testing)
test_x, test_y = neuralNet.split(test)
res_y = model.predict(test_x)
l = []
for i in range(len(res_y[0])):
pred = np.array([])
correct = np.array([])
for j in range(len(test_y)):
import utils
import save_arrays
import sys
if __name__ == '__main__':
name = "goals"
read_folder, save_path = utils.get_folders(sys.argv, name)
plans = utils.get_plans(read_folder)
goals = []
for p in plans:
if p.goals not in goals:
goals.append(p.goals)
save_arrays.save(goals, save_path)
print(len(goals))
dictionary[num_actions[i]] = 1
else:
dictionary[num_actions[i]] += 1
dictionary = {
k: v
for k, v in sorted(
dictionary.items(), key=lambda item: item[0], reverse=False)
}
vals = np.fromiter(dictionary.values(), dtype=float)
file = open("num_azioni_nei_piani.txt", "w")
for k, w in dictionary.items():
file.write("Numero azioni nel piano: " + str(k) + " Frequenza: " +
str(w) + "\n")
file.close()
plt.plot(vals)
plt.xlabel('numero_azioni_nei_piani')
plt.ylabel('frequenza_numero_azioni')
plt.savefig("plot_distribuzione_numero_azioni_nei_piani.png")
plt.show()
if __name__ == '__main__':
folder = "XmlPlans"
plans = get_plans(folder)
plot_num_actions_plans(plans)
domains = logistics_domains(folder)
init_statistics(domains)
save_arrays.save(domains, "./files/domains.obj")
save_arrays.save(plans, "./files/piani.obj")
encoder(domains)
def save_db(db):
np.random.shuffle(db)
dim = int(0.8 * len(db))
train, test = db[:dim], db[dim:]
save_arrays.save(train, "training_set")
save_arrays.save(test, "test_set")
import utils
import numpy as np
import save_arrays
import pickle
import oneHot_deep
if __name__ == '__main__':
file = open("goal", "rb")
goal = pickle.load(file)
dizionario_goal = oneHot_deep.create_dictionary_goals(goal)
dizionario_goal = oneHot_deep.shuffle_dictionary(dizionario_goal)
oneHot_deep.completa_dizionario(dizionario_goal)
save_arrays.save(dizionario_goal, "dizionario_goal")
import crea_istanze
if __name__ == '__main__':
file1 = open("apn.obj", "rb")
apn_list = pickle.load(file1)
file2 = open("cit.obj", "rb")
cit_list = pickle.load(file2)
file3 = open("obj.obj", "rb")
obj_list = pickle.load(file3)
file4 = open("loc.obj", "rb")
loc_list = pickle.load(file4)
file5 = open("tru.obj", "rb")
tru_list = pickle.load(file5)
apn_list = na.natsorted(apn_list[1:])
cit_list = na.natsorted(cit_list)
obj_list = na.natsorted(obj_list[1:])
tru_list = na.natsorted(tru_list)
loc_list = na.natsorted(loc_list)
folder = "SOL_files"
plans = utils.get_plans(folder)
oneHot.init(plans, apn_list, cit_list, obj_list, loc_list, tru_list)
db = crea_istanze.crea(plans)
np.random.shuffle(db)
dim = int(0.8 * len(db))
train, test = db[:dim], db[dim:]
save_arrays.save(train, "training_set")
save_arrays.save(test, "test_set")
print('dataset created')