def test_containsOutliers5D(self):
d = 5
cluster_A = Cluster([(0.0, 3.0, 0.0, 2.0, 0.0),
(0.0, 4.0, 0.0, 2.0, 0.0),
(0.0, 4.0, 0.0, 1.0, 0.0),
(0.0, 3.0, 0.0, 1.0, 0.0)], d)
cluster_B = Cluster([(0.0, 3.0, 0.0, 5.0, 0.0),
(0.0, 5.0, 0.0, 5.0, 0.0),
(0.0, 4.0, 0.0, 6.0, 0.0)], d)
samples = SampleContainer([(0.0, 7.0, 0.0, 5.0, 0.0),
(0.0, 7.0, 0.0, 4.0, 0.0)], d)
resul = containsOutlier(mergeClusters(cluster_A, cluster_B), samples)
self.assertEqual(
False, resul,
"Dos clusters deberian ser mergeables si forman una componente convexa sin outliers en dimension %s"
% (d))
samples = SampleContainer([(0.0, 4.0, 0.0, 4.0, 0.0),
(0.0, 3.0, 0.0, 4.0, 0.0),
(0.0, 5.0, 0.0, 4.0, 0.0)], d)
resul = containsOutlier(mergeClusters(cluster_A, cluster_B), samples)
self.assertEqual(
True, resul,
"Dos clusters no deberian ser mergeables si forman una componente convexa con outliers %s"
% (d))
def test_merge_2D(self):
d = 2
cluster_A = Cluster([(0.0, 0.0), (1.0, 1.0)], d)
cluster_B = Cluster([(2.0, 2.0), (3.0, 3.0)], d)
c_test = Cluster([(0.0, 0.0), (1.0, 1.0), (2.0, 2.0), (3.0, 3.0)], d)
c = mergeClusters(cluster_A, cluster_B)
self.assertEquals(c_test, c, "los cluster deben ser iguales")
self.assertNotEquals(cluster_A, c, "los clusters no son iguales")
self.assertNotEquals(cluster_B, c, "los clusters no son iguales")
def test_containsOutliers2D(self):
d = 2
cluster_A = Cluster([(0.0, 2.0), (0.0, 4.0)], d)
cluster_B = Cluster([(4.0, 2.0), (4.0, 4.0)], d)
samples = SampleContainer([(6.0, 3.0)], d)
resul = containsOutlier(mergeClusters(cluster_A, cluster_B), samples)
self.assertEqual(
False, resul,
"Dos clusters deberian ser mergeables si forman una componente convexa sin outliers"
)
samples = SampleContainer([(2.3, 3.0)], d)
resul = containsOutlier(mergeClusters(cluster_A, cluster_B), samples)
self.assertEqual(
True, resul,
"Dos clusters no deberian ser mergeables si forman una componente convexa con outliers"
)
pass
def testisMergeableEmptyOutliers(self):
d = 2
cluster_A = Cluster([(0.0, 2.0), (0.0, 4.0)], d)
cluster_B = Cluster([(4.0, 2.0), (4.0, 4.0)], d)
samples = SampleContainer([], d)
resul = containsOutlier(mergeClusters(cluster_A, cluster_B), samples)
self.assertEqual(
False, resul,
"Dos clusters deberian ser mergeables si no hay otras muestras")
def createClusters(samplesA, samplesB):
clusters = None
if not isinstance(samplesA, ClusterContainer):
print("no es sample")
clusters = createDefaultClusters(samplesA)
else:
print("si es sample")
clusters = samplesA
samples = samplesB
if samplesA.getSize() == 1:
return clusters
else:
K = clusters.getSize()
k = 0
distances_graph = createDistanceGraph(clusters.getClusters())
#sorted_edges = sorted(distances_graph.edges(data=True), key=lambda x: x[2]['weight'])
while k < K:
#(u,v,w) = sorted_edges[0]
(u, v) = minimumEdge(distances_graph)
merged = mergeClusters(u, v)
"""print("se puede fusionar: " + str(not containsOutlier(merged, samples)) + " k: " + str(k) + " K: " + str(K))
print("cluster u: " + str(map(lambda s : s.getData(), u.getSamples())))
print("cluster v: " + str(map(lambda s : s.getData(), v.getSamples())))
print("cluster merged: " + str(map(lambda s : s.getData(), merged.getSamples())))"""
if containsOutlier(merged, samples):
#k = k + 1
#sorted_edges.remove(sorted_edges[0])
distances_graph[u][v]['weight'] = float('inf')
else:
clusters = updateClusterContainer(clusters, u, v, merged)
distances_graph = updateDistanceGraph(distances_graph, u, v,
merged)
#sorted_edges = sorted(distances_graph.edges(data=True), key=lambda x: x[2]['weight'])
K = K - 1
k = 0
k = k + 1
return ClusterContainer(
filter(lambda cls: cls.getSize() >= samplesA.getSize() * 0.01,
clusters.getClusters()), clusters.getDimension())
def createClusters2(samplesA, samplesB):
clusters = createDefaultClusters(samplesA)
samples = samplesB
if samplesA.getSize() == 1:
return clusters
else:
K = clusters.getSize()
k = 0
print("creando grafo de distancias...")
distances_graph = createDistanceGraph(clusters.getClusters())
#print("ordenando aristas...")
sorted_edges = sorted(distances_graph.edges(data=True),
key=lambda x: x[2]['weight'])
has_already_been_merged = createMap(distances_graph.nodes)
#print("Cantidad de aristas " + str(len(sorted_edges)))
#print("Cantidad de clusters: " + str(clusters.getSize()))
print("reduciendo clusters...")
while k < K:
if len(sorted_edges) == 0:
print("re-ordenando...")
sorted_edges = sorted(distances_graph.edges(data=True),
key=lambda x: x[2]['weight'])
has_already_been_merged = createMap(distances_graph.nodes)
"""print("Cantidad de aristas " + str(len(sorted_edges)))
print("Cantidad de clusters: " + str(clusters.getSize()))
print("cantidad de clusters: " + str(map(lambda s: s.getSize(),clusters.getClusters())))
print("K: " + str(K))"""
else:
(u, v, w) = sorted_edges[0]
#(u,v) = minimumEdge(distances_graph)
if (not has_already_been_merged[v]
and not has_already_been_merged[u]):
merged = mergeClusters(u, v)
"""print("se puede fusionar: " + str(not containsOutlier(merged, samples)) + " k: " + str(k) + " K: " + str(K))
print("cluster u: " + str(map(lambda s : s.getData(), u.getSamples())))
print("cluster v: " + str(map(lambda s : s.getData(), v.getSamples())))
print("cluster merged: " + str(map(lambda s : s.getData(), merged.getSamples())))"""
if not containsOutlier(merged, samples):
clusters = updateClusterContainer(
clusters, u, v, merged)
distances_graph = updateDistanceGraph(
distances_graph, u, v, merged)
has_already_been_merged[v] = True
has_already_been_merged[u] = True
K = K - 1
k = 0
k = k + 1
sorted_edges.remove(sorted_edges[0])
print(map(lambda c: c.getSize(), clusters.getClusters()))
clusters = createClusters(clusters, samplesB)
return ClusterContainer(
filter(lambda cls: cls.getSize() >= samplesA.getSize() * 0.01,
clusters.getClusters()), clusters.getDimension())
