def test_Point2D():
point_int = cl.Point2D(3, 4)
point_float = cl.Point2D(3.5, 2.1)
errors = []
if not (point_int.X == 3 and point_int.Y == 4):
errors.append('error')
if not (point_float.X == 3.5 and point_float.Y == 2.1):
errors.append('error')
if not (point_float.__repr__() == '(3.50, 2.10)' and point_int.__repr__() == '(3.00, 4.00)'):
errors.append('error')
if not point_int + point_float == cl.Point2D(6.5, 6.1):
errors.append("error")
if not point_float - point_int == cl.Point2D(0.5, -1.9):
errors.append("error")
if not point_float * point_int == cl.Point2D(10.5, 8.4):
errors.append("error")
point_int += point_float
if not point_int == cl.Point2D(6.5, 6.1):
errors.append("error")
point_int -= point_float
if point_int != cl.Point2D(3, 4):
pass
#errors.append("error")
point_int *= point_float
if not point_int == cl.Point2D(10.5, 8.4):
pass
#errors.append("error")
assert not errors
def test_turnLeft():
points_set = [
classes.Point2D(1, 1),
classes.Point2D(2, 2),
classes.Point2D(2.5, 3),
classes.Point2D(3, 2.5),
classes.Point2D(3, 3)
]
errors = []
if not utils.turnLeft(points_set[:2], points_set[2]):
errors.append("error")
if utils.turnLeft(points_set[:2], points_set[3]):
errors.append("error")
if utils.turnLeft(points_set[:2], points_set[4]):
errors.append("error")
assert not errors
def move(self, W, C1, C2, Gb):
R1, R2 = ranf(size=2)
for i in xrange(len(self.centroids)):
self.vel[i] = (W * self.vel[i]) + (C1 * R1 * (utils.euclidean(self.best_centroids[i], self.centroids[i])))\
+ (C2 * R2 * (utils.euclidean(Gb[i] , self.centroids[i])))
self.centroids[i] += classes.Point2D(*self.vel[i])
self.recluster()
if self.use_variance:
self.fit_by_variance()
else:
self.fit_by_area()
if self.fitness > self.best_fitness:
self.best_fitness = self.fitness
self.best_centroids = self.centroids[:]
def generate_population(self, n_clusters=13):
population = []
for _ in tqdm(xrange(self.n_particles),
desc='Generando Poblacion inicial',
unit=' Particle'):
individuo = []
cluster = KMeans(n_clusters=n_clusters, random_state=self.seed)
cluster_labels = cluster.fit_predict(self.data[:, 0:2])
individuo_df = pd.DataFrame({
'x': self.data[:, 0],
'y': self.data[:, 1],
'time_store': self.data[:, 2],
'cluster': cluster_labels
})
for i in xrange(n_clusters):
individuo.append(
classes.ClusterPdV([
classes.PdV(*point[:-1]) for point in individuo_df[
individuo_df.cluster == i].values
], classes.Point2D(*cluster.cluster_centers_[i])))
population.append(Particle(individuo, self.use_var))
return population
def test_getConvexHull():
errors = []
points_set = [
classes.Point2D(2.5, 0),
classes.Point2D(3.5, 0.5),
classes.Point2D(1, 1),
classes.Point2D(2, 2),
classes.Point2D(3, 2.5),
classes.Point2D(4, 3),
classes.Point2D(2.5, 3),
classes.Point2D(3, 3)
]
convex_hull = utils.getConvexHull(points_set)
if not convex_hull == [
classes.Point2D(1, 1),
classes.Point2D(2.5, 3),
classes.Point2D(4, 3),
classes.Point2D(3.5, 0.5),
classes.Point2D(2.5, 0)
]:
errors.append("error")
if not points_set != [
classes.Point2D(2.5, 0),
classes.Point2D(3.5, 0.5),
classes.Point2D(1, 1),
classes.Point2D(2, 2),
classes.Point2D(3, 2.5),
classes.Point2D(4, 3),
classes.Point2D(2.5, 3),
classes.Point2D(3, 3)
]:
errors.append("error")
assert not errors
def test_sortByX():
a = [
classes.Point2D(1, 1),
classes.Point2D(2, 2),
classes.Point2D(3, 3),
classes.Point2D(3, 2.5),
classes.Point2D(2.5, 3)
]
errors = []
random.shuffle(a)
utils.sortByX(a)
if not a == [
classes.Point2D(1, 1),
classes.Point2D(2, 2),
classes.Point2D(2.5, 3),
classes.Point2D(3, 2.5),
classes.Point2D(3, 3)
]:
errors.append('error')
random.shuffle(a)
utils.sortByX(a, reverse=True)
if not a == [
classes.Point2D(3, 3),
classes.Point2D(3, 2.5),
classes.Point2D(2.5, 3),
classes.Point2D(2, 2),
classes.Point2D(1, 1)
]:
errors.append('error')
assert not errors