def test_set_case(self, process):
assert process.check_point_cases == 0
assert process.cases == []
process.set_case('5', 'activityA', datetime(2015, 5, 10, 8, 00, 00))
assert process.check_point_cases == 1
assert process.cases
case_test = process.cases[0]
assert case_test.id == '5'
activity = case_test.activities[0]
assert activity.name == 'activityA'
assert activity.timestamp == datetime(2015, 5, 10, 8, 00, 00)
case1 = Case('1')
case2 = Case('2')
case3 = Case('3')
case4 = Case('4')
process.check_point_cases = 0
process.cases = [case1, case2, case3, case4]
process.set_case('3', 'activityA', datetime(2015, 5, 10, 8, 00, 00))
assert process.check_point_cases == 0
assert process.cases
case_test = process.cases[0]
assert case_test.id == '3'
activity = case_test.activities[0]
assert activity.name == 'activityA'
assert activity.timestamp == datetime(2015, 5, 10, 8, 00, 00)
def cases_list(self):
case_list = []
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_list.append(case)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
case_list.append(case)
case = Case('4')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case)
return case_list
def simple_graph(self):
case_list = []
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_list.append(case)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
case_list.append(case)
graph = nx.DiGraph(name='simple_graph')
graph = initialize_graph(graph, case_list)
return graph
def test_extract_cases_time_and_trace():
case_list = []
with pytest.raises(Exception):
assert extract_cases_time_and_trace(case_list)
case = Case('1')
case.set_activity('activity1', datetime(2015, 5, 10, 8, 30, 00))
case.set_activity('activity2', datetime(2015, 5, 10, 9, 00, 00))
case.set_activity('activity3', datetime(2015, 5, 10, 9, 30, 00))
case.set_activity('activity4', datetime(2015, 5, 10, 10, 00, 00))
case_list.append(case)
case = Case('2')
case.set_activity('activity1', datetime(2015, 5, 10, 8, 5, 30))
case.set_activity('activity3', datetime(2015, 5, 10, 9, 33, 52))
case.set_activity('activity2', datetime(2015, 5, 10, 10, 30, 00))
case.set_activity('activity4', datetime(2015, 5, 10, 11, 50, 00))
case_list.append(case)
case = Case('3')
case.set_activity('activity5', datetime(2016, 5, 10, 8, 5, 30))
case.set_activity('activity1', datetime(2017, 5, 10, 9, 33, 52))
case.set_activity('activity3', datetime(2018, 5, 10, 10, 30, 00))
case.set_activity('activity4', datetime(2019, 5, 10, 11, 50, 00))
case_list.append(case)
trace_list, time_list = extract_cases_time_and_trace(case_list)
expected_trace_list = [[
'activity1', 'activity2', 'activity3', 'activity4'
], ['activity1', 'activity3', 'activity2',
'activity4'], ['activity5', 'activity1', 'activity3', 'activity4']]
expected_time_list = [[
datetime(2015, 5, 10, 8, 30, 00),
datetime(2015, 5, 10, 9, 00, 00),
datetime(2015, 5, 10, 9, 30, 00),
datetime(2015, 5, 10, 10, 00, 00)
],
[
datetime(2015, 5, 10, 8, 5, 30),
datetime(2015, 5, 10, 9, 33, 52),
datetime(2015, 5, 10, 10, 30, 00),
datetime(2015, 5, 10, 11, 50, 00)
],
[
datetime(2016, 5, 10, 8, 5, 30),
datetime(2017, 5, 10, 9, 33, 52),
datetime(2018, 5, 10, 10, 30, 00),
datetime(2019, 5, 10, 11, 50, 00)
]]
assert len(trace_list) == 3
assert len(time_list) == 3
assert trace_list == expected_trace_list
assert time_list == expected_time_list
def simple_graph(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
})
case_list.append(case)
graph = initialize_graph(case_list)
return graph
def test_calculate_empty_distances(self):
empty_case = Case("Empty case")
assert calculate_case_time_distances(empty_case) == [0]
single_event_case = Case("Single event case")
single_event_case.add_event({
"concept:name":
"",
"time:timestamp":
datetime(2015, 5, 10, 8, 30, 30)
})
assert calculate_case_time_distances(single_event_case) == [0]
def test_initialize_cdesf(self, process, cases_list):
assert process.nyquist == 0
assert process.check_point_cases == 0
assert len(process.process_model_graph.edges) == 0
assert len(process.process_model_graph.nodes) == 0
assert process.cases == []
with pytest.raises(Exception):
assert initialize_graph([])
case4 = Case("4")
case4.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case4.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case4.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
process.cases = cases_list
process.cases.insert(0, case4)
process.process_model_graph = initialize_graph(process.cases)
pmg = initialize_graph(cases_list)
for case in cases_list:
case.distances = calculate_case_distances(pmg, case)
def test_generate_outlier_clusters(self, denstream):
assert denstream.generate_outlier_clusters() == []
case_0 = Case("0")
case_1 = Case("1")
case_2 = Case("2")
case_3 = Case("3")
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.weight = 3
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.weight = 4
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(3, 2, 0, 0.15)
micro_cluster.weight = 6
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
assert denstream.all_cases == {"0": 0, "1": 1, "2": 3, "3": 3}
cluster_list = denstream.generate_outlier_clusters()
cluster = cluster_list[0]
assert cluster.id == 0
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 3
assert cluster.case_ids == ["0"]
cluster = cluster_list[1]
assert cluster.id == 1
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 4
assert cluster.case_ids == ["1"]
cluster = cluster_list[2]
assert cluster.id == 3
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 6
assert cluster.case_ids == ["2", "3"]
def test_generate_outlier_clusters(self, denstream):
assert denstream.generate_outlier_clusters() == []
case_0 = Case('0')
case_1 = Case('1')
case_2 = Case('2')
case_3 = Case('3')
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.weight = 3
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.weight = 4
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(3, 2, 0, 0.15)
micro_cluster.weight = 6
denstream.o_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
assert denstream.all_cases == {'0': 0, '1': 1, '2': 3, '3': 3}
cluster_list = denstream.generate_outlier_clusters()
cluster = cluster_list[0]
assert cluster.id == 0
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 3
assert cluster.case_ids == ['0']
cluster = cluster_list[1]
assert cluster.id == 1
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 4
assert cluster.case_ids == ['1']
cluster = cluster_list[2]
assert cluster.id == 3
assert np.all(cluster.centroid == [0, 0])
assert cluster.radius == 0
assert cluster.weight == 6
assert cluster.case_ids == ['2', '3']
def test_update():
case_list = []
micro_cluster = MicroCluster(0, 2, 0, 0.15)
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 33, 20))
case.set_activity('activityE', datetime(2015, 5, 10, 14, 6, 40))
case_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 1, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 14, 40, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 15, 5, 00))
case_list.append(case)
graph = nx.DiGraph()
graph = initialize_graph(graph, case_list)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 13, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 13, 00))
case.graph_distance, case.time_distance = extract_case_distances(
graph, case)
cf = micro_cluster.CF.copy()
cf2 = micro_cluster.CF2.copy()
weight = micro_cluster.weight
micro_cluster.update(case)
assert np.all(micro_cluster.CF == cf + case.point)
assert np.all(micro_cluster.CF2 == cf2 + case.point * case.point)
assert micro_cluster.weight == weight + 1
case = case_list[0]
case.graph_distance, case.time_distance = extract_case_distances(
graph, case)
cf = micro_cluster.CF.copy()
cf2 = micro_cluster.CF2.copy()
weight = micro_cluster.weight
micro_cluster.update(case)
assert np.all(micro_cluster.CF == cf + case.point)
assert np.all(micro_cluster.CF2 == cf2 + case.point * case.point)
assert micro_cluster.weight == weight + 1
def test_get_case(self, process):
case1 = Case('1')
case2 = Case('2')
case3 = Case('3')
case4 = Case('4')
process.cases = [case1, case2, case3, case4]
case = process.get_case('1')
assert case == 0
case = process.get_case('3')
assert case == 2
process.cases = [case4, case1, case3, case2]
case = process.get_case('1')
assert case == 1
case = process.get_case('2')
assert case == 3
case = process.get_case('5')
assert case is None
def test_get_case(self, process):
case1 = Case("1")
case2 = Case("2")
case3 = Case("3")
case4 = Case("4")
process.cases = [case1, case2, case3, case4]
case = process.get_case_index("1")
assert case == 0
case = process.get_case_index("3")
assert case == 2
process.cases = [case4, case1, case3, case2]
case = process.get_case_index("1")
assert case == 1
case = process.get_case_index("2")
assert case == 3
case = process.get_case_index("5")
assert case is None
def test_set_case(self, process: CDESF):
assert process.check_point_cases == 0
assert process.cases == []
process.set_case(
"5",
{
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
},
)
assert process.check_point_cases == 1
assert process.cases
case_test = process.cases[0]
assert case_test.id == "5"
assert case_test.events[0].get("concept:name") == "activityA"
assert case_test.events[0].get("time:timestamp") == datetime(
2015, 5, 10, 8, 00, 00)
case1 = Case("1")
case2 = Case("2")
case3 = Case("3")
case4 = Case("4")
process.check_point_cases = 0
process.cases = [case1, case2, case3, case4]
process.set_case(
"3",
{
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
},
)
assert process.check_point_cases == 0
assert process.cases
case_test = process.cases[0]
assert case_test.id == "3"
assert case_test.events[0].get("concept:name") == "activityA"
assert case_test.events[0].get("time:timestamp") == datetime(
2015, 5, 10, 8, 00, 00)
def test_initialize_cdesf(self, process, cases_list):
assert process.nyquist == 0
assert process.check_point_cases == 0
assert len(process.process_model_graph.edges) == 0
assert len(process.process_model_graph.nodes) == 0
assert process.cases == []
with pytest.raises(Exception):
assert initialize_graph(nx.DiGraph(), [])
case4 = Case('4')
case4.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case4.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case4.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
process.cases = cases_list
process.cases.insert(0, case4)
process.process_model_graph = initialize_graph(nx.DiGraph(), process.cases)
pmg = initialize_graph(nx.DiGraph(), cases_list)
for case in cases_list:
graph_distance, time_distance = extract_case_distances(pmg, case)
case.graph_distance = graph_distance
case.time_distance = time_distance
case_4 = process.cases[0]
process.initialize_cdesf()
graph_dist_4, time_dist_4 = extract_case_distances(pmg, case_4)
assert case_4.graph_distance == graph_dist_4
assert case_4.time_distance == time_dist_4
assert len(process.denstream.p_micro_clusters) == 1
assert process.denstream.all_cases.keys() == {'1', '4'}
new_p_mc = process.denstream.p_micro_clusters[0]
assert new_p_mc.weight == 2
assert new_p_mc.creation_time == 0
assert new_p_mc.lambda_ == 0.15
assert process.initialized
def test_release_cases_from_memory(self, process):
case1 = Case('1')
case2 = Case('2')
case3 = Case('3')
case4 = Case('4')
case4.set_activity('activity1', datetime(2015, 5, 10, 8, 30, 00))
case1.set_activity('activity1', datetime(2015, 5, 10, 9, 00, 00))
case3.set_activity('activity1', datetime(2015, 5, 10, 9, 30, 00))
case2.set_activity('activity1', datetime(2015, 5, 10, 10, 00, 00))
process.cases = [case2, case3, case1, case4]
process.release_cases_from_memory()
assert process.cases == []
process.cases = [case2, case3, case1, case4]
process.nyquist = 1
process.release_cases_from_memory()
assert process.cases == [case2]
process.cases = [case2, case3, case1, case4]
process.nyquist = 2
process.release_cases_from_memory()
assert process.cases == [case2, case3]
process.cases = [case2, case3, case1, case4]
process.nyquist = 3
process.release_cases_from_memory()
assert process.cases == [case2, case3, case1]
process.cases = [case2, case3, case1, case4]
process.nyquist = 4
process.release_cases_from_memory()
assert process.cases == [case2, case3, case1, case4]
def test_train(self, denstream):
#anomalous_denstream = DenStream(2, 0.15, 0.3, 0.1, 4, 1000)
cases_list = []
case = Case('5')
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
case.set_activity('activityE', datetime(2015, 5, 10, 8, 00, 40))
cases_list.append(case)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
cases_list.append(case)
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
cases_list.append(case)
case = Case('4')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
cases_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
cases_list.append(case)
pmg = initialize_graph(nx.DiGraph(), cases_list)
for case in cases_list:
graph_distance, time_distance = extract_case_distances(pmg, case)
case.graph_distance = graph_distance
case.time_distance = time_distance
denstream.dbscan(cases_list)
assert denstream.stream_speed == 1000
assert denstream.no_processed_points == 0
assert denstream.time == 0
assert len(denstream.p_micro_clusters) == 1
assert len(denstream.o_micro_clusters) == 0
case_5 = cases_list[0]
denstream.train(case_5)
assert denstream.all_cases == {'1': 0, '4': 0, '5': 1}
assert denstream.time == 0
assert denstream.no_processed_points == 1
assert not denstream.p_micro_clusters == []
assert not denstream.o_micro_clusters == []
assert len(denstream.p_micro_clusters) == 1
assert len(denstream.o_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [0.25, 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [0.03125, 0])
assert denstream.p_micro_clusters[0].weight == 2
assert denstream.p_micro_clusters[0].creation_time == 0
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
case_3 = cases_list[1]
denstream.train(case_3)
assert denstream.all_cases == {'1': 0, '3': 0, '4': 0, '5': 1}
assert denstream.time == 0
assert denstream.no_processed_points == 2
assert len(denstream.p_micro_clusters) == 1
assert len(denstream.o_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [0.5, 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [0.09375, 0])
assert denstream.p_micro_clusters[0].weight == 3
assert denstream.p_micro_clusters[0].creation_time == 0
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
case_1 = cases_list[2]
denstream.train(case_1)
assert denstream.all_cases == {'1': 0, '3': 0, '4': 0, '5': 1}
assert denstream.time == 0
assert denstream.no_processed_points == 3
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [(7 / 64), 0])
assert denstream.p_micro_clusters[0].weight == 4
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
case_4 = cases_list[3]
denstream.train(case_4)
assert denstream.all_cases == {'1': 0, '3': 0, '4': 0, '5': 1}
assert denstream.time == 0
assert denstream.no_processed_points == 4
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [0.75, 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [0.125, 0])
assert denstream.p_micro_clusters[0].weight == 5
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
case_1 = cases_list[4]
denstream.train(case_1)
assert denstream.all_cases == {'1': 0, '2': 0, '3': 0, '4': 0, '5': 1}
assert denstream.time == 0
assert denstream.no_processed_points == 5
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [0.75, 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [0.125, 0])
assert denstream.p_micro_clusters[0].weight == 6
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
case = Case('6')
case.set_activity('activityE', datetime(2015, 5, 10, 8, 00, 40))
case.set_activity('activityF', datetime(2015, 5, 10, 8, 00, 50))
graph_distance, time_distance = extract_case_distances(pmg, case)
case.graph_distance = graph_distance
case.time_distance = time_distance
cases_list.insert(0, case)
denstream.train(cases_list[0])
assert denstream.all_cases == {
'1': 0,
'2': 0,
'3': 0,
'4': 0,
'5': 1,
'6': 2
}
assert denstream.time == 0
assert denstream.no_processed_points == 6
assert len(denstream.o_micro_clusters) == 2
assert len(denstream.p_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == [0.75, 0])
assert np.all(denstream.p_micro_clusters[0].CF2 == [0.125, 0])
assert denstream.p_micro_clusters[0].weight == 6
assert denstream.o_micro_clusters[0].id == 1
assert np.all(denstream.o_micro_clusters[0].CF == [(5 / 8), 0])
assert np.all(denstream.o_micro_clusters[0].CF2 == [(5 / 8)**2, 0])
assert denstream.o_micro_clusters[0].weight == 1
assert denstream.o_micro_clusters[0].creation_time == 0
assert denstream.o_micro_clusters[1].id == 2
assert np.all(denstream.o_micro_clusters[1].CF == [1, 1])
assert np.all(denstream.o_micro_clusters[1].CF2 == [1, 1])
assert denstream.o_micro_clusters[1].weight == 1
assert denstream.o_micro_clusters[1].creation_time == 0
assert denstream.all_cases == {
'1': 0,
'2': 0,
'3': 0,
'4': 0,
'5': 1,
'6': 2
}
denstream.stream_speed = 7
denstream.train(cases_list[0])
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 2
CF = np.array([0.75, 0]) * (2**(-0.15))
CF2 = np.array([0.125, 0]) * (2**(-0.15))
weight = 6 * (2**(-0.15))
assert denstream.p_micro_clusters[0].id == 0
assert np.all(denstream.p_micro_clusters[0].CF == CF)
assert np.all(denstream.p_micro_clusters[0].CF2 == CF2)
assert denstream.p_micro_clusters[0].weight == weight
CF = (np.array([1, 1]) + cases_list[0].point) * (2**(-0.15))
CF2 = (np.array([1, 1]) +
(cases_list[0].point * cases_list[0].point)) * (2**(-0.15))
weight = (1 + 1) * (2**(-0.15))
assert denstream.p_micro_clusters[1].id == 2
assert np.all(denstream.p_micro_clusters[1].CF == CF)
assert np.all(denstream.p_micro_clusters[1].CF2 == CF2)
assert denstream.p_micro_clusters[1].weight == weight
for i in range(16):
assert denstream.all_cases == {
'1': 0,
'2': 0,
'3': 0,
'4': 0,
'5': 1,
'6': 2
}
denstream.stream_speed = 8 + i
# mc.weight > self.beta * self.mu
denstream.train(cases_list[0])
assert denstream.time == i + 2
assert denstream.p_micro_clusters[0].id == 0
CF = np.array([0.75, 0]) * (2**(-0.15))
CF2 = np.array([0.125, 0]) * (2**(-0.15))
weight = 6 * (2**(-0.15))
for x in range(i + 1):
CF *= (2**(-0.15))
CF2 *= (2**(-0.15))
weight *= (2**(-0.15))
assert np.all(denstream.p_micro_clusters[0].CF == CF)
assert np.all(denstream.p_micro_clusters[0].CF2 == CF2)
assert denstream.p_micro_clusters[0].weight == weight
assert denstream.p_micro_clusters[1].id == 2
CF_1 = (np.array([1, 1]) + cases_list[0].point) * (2**(-0.15))
CF2_1 = (np.array([1, 1]) +
(cases_list[0].point * cases_list[0].point)) * (2
**(-0.15))
weight_1 = (1 + 1) * (2**(-0.15))
for x in range(i + 1):
CF_1 += cases_list[0].point
CF_1 *= (2**(-0.15))
CF2_1 += cases_list[0].point * cases_list[0].point
CF2_1 *= (2**(-0.15))
weight_1 += 1
weight_1 *= (2**(-0.15))
assert np.all(denstream.p_micro_clusters[1].CF == CF_1)
assert np.all(denstream.p_micro_clusters[1].CF2 == CF2_1)
assert denstream.p_micro_clusters[1].weight == weight_1
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 2
assert denstream.time == 17
assert denstream.no_processed_points == 23
assert denstream.stream_speed == 23
denstream.stream_speed += 1
case = Case('7')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityG', datetime(2015, 5, 10, 8, 1, 00))
graph_distance, time_distance = extract_case_distances(pmg, case)
case.graph_distance = graph_distance
case.time_distance = time_distance
cases_list.insert(0, case)
assert cases_list[0].id == '7'
denstream.train(cases_list[0])
assert denstream.all_cases == {
'1': 0,
'2': 0,
'3': 0,
'4': 0,
'5': 1,
'6': 2,
'7': 3
}
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 1
assert denstream.p_micro_clusters[0].id == 2
CF_1 *= (2**(-0.15))
CF2_1 *= (2**(-0.15))
weight_1 *= (2**(-0.15))
assert np.all(denstream.p_micro_clusters[0].CF == CF_1)
assert np.all(denstream.p_micro_clusters[0].CF2 == CF2_1)
assert denstream.p_micro_clusters[0].weight == weight_1
assert denstream.o_micro_clusters[0].id == 3
assert np.all(denstream.o_micro_clusters[0].CF == [1, log10(60)])
assert np.all(denstream.o_micro_clusters[0].CF2 == [1, log10(60)**2])
assert denstream.o_micro_clusters[0].weight == 1
def test_generate_clusters(self, denstream):
assert len(denstream.p_micro_clusters) == 0
dense_group, not_dense_group = denstream.generate_clusters()
assert dense_group == [[]]
assert not_dense_group == [[]]
case_0 = Case('0')
case_1 = Case('1')
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([5.0, 5.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 5
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
assert len(denstream.p_micro_clusters) == 1
assert micro_cluster.weight >= denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert not dense_group == [[]]
assert not_dense_group == [[]]
cluster_list = dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 5
assert cluster.case_ids == ['0', '1']
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([3.0, 3.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 3
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
assert len(denstream.p_micro_clusters) == 1
assert micro_cluster.weight < denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert dense_group == [[]]
assert not not_dense_group == [[]]
cluster_list = not_dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 3
assert cluster.case_ids == ['0', '1']
case_2 = Case('2')
case_3 = Case('3')
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([3.0, 3.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 3
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([4.0, 4.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 4
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) <= 2 * denstream.epsilon
assert cl1.weight + cl2.weight >= denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert not dense_group == [[]]
assert not_dense_group == [[]]
cluster_list = dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 3
assert cluster.case_ids == ['0', '1']
cluster = cluster_list[1]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 4
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert dense_group[1]
assert cluster_list[2]
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([3.0, 3.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 1
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([3.0, 3.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 1
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) <= 2 * denstream.epsilon
assert cl1.weight + cl2.weight < denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert dense_group == [[]]
assert not not_dense_group == [[]]
cluster_list = not_dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [3, 3])
assert cluster.radius == 0
assert cluster.weight == 1
assert cluster.case_ids == ['0', '1']
cluster = cluster_list[1]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [3, 3])
assert cluster.radius == 0
assert cluster.weight == 1
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert not_dense_group[1]
assert cluster_list[2]
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([3.0, 3.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 3
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([1.0, 1.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 4
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) > 2 * denstream.epsilon
assert cl1.weight < denstream.mu
assert cl2.weight >= denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert not dense_group == [[]]
assert not not_dense_group == [[]]
cluster_list = not_dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 3
assert cluster.case_ids == ['0', '1']
with pytest.raises(IndexError):
assert cluster_list[1]
cluster_list = dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [0.25, 0.25])
assert cluster.radius == cl2.radius
assert cluster.weight == 4
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert not_dense_group[1]
assert dense_group[1]
assert cluster_list[1]
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([6.0, 6.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 6
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([1.0, 1.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 2
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) > 2 * denstream.epsilon
assert cl1.weight >= denstream.mu
assert cl2.weight < denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert not dense_group == [[]]
assert not not_dense_group == [[]]
cluster_list = dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 6
assert cluster.case_ids == ['0', '1']
with pytest.raises(IndexError):
assert cluster_list[1]
cluster_list = not_dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [0.5, 0.5])
assert cluster.radius == cl2.radius
assert cluster.weight == 2
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert not_dense_group[1]
assert dense_group[1]
assert cluster_list[1]
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([6.0, 6.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 6
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([1.0, 1.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 4
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) > 2 * denstream.epsilon
assert cl1.weight >= denstream.mu
assert cl2.weight >= denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert not dense_group == [[]]
assert not_dense_group == [[]]
cluster_list = dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [1, 1])
assert cluster.radius == 0
assert cluster.weight == 6
assert cluster.case_ids == ['0', '1']
with pytest.raises(IndexError):
assert cluster_list[1]
cluster_list = dense_group[1]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [0.25, 0.25])
assert cluster.radius == cl2.radius
assert cluster.weight == 4
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert dense_group[2]
assert cluster_list[1]
micro_cluster = MicroCluster(0, 2, 0, 0.15)
micro_cluster.CF = np.array([1.0, 1.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 3
denstream.p_micro_clusters = []
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_0.id] = micro_cluster.id
denstream.all_cases[case_1.id] = micro_cluster.id
micro_cluster = MicroCluster(1, 2, 0, 0.15)
micro_cluster.CF = np.array([1.0, 1.0])
micro_cluster.CF2 = np.array([1.0, 1.0])
micro_cluster.weight = 2
denstream.p_micro_clusters.append(micro_cluster)
denstream.all_cases[case_2.id] = micro_cluster.id
denstream.all_cases[case_3.id] = micro_cluster.id
cl1 = denstream.p_micro_clusters[0]
cl2 = denstream.p_micro_clusters[1]
assert len(denstream.p_micro_clusters) > 1
assert denstream.euclidean_distance(
cl1.centroid, cl2.centroid) > 2 * denstream.epsilon
assert cl1.weight < denstream.mu
assert cl2.weight < denstream.mu
dense_group, not_dense_group = denstream.generate_clusters()
assert dense_group == [[]]
assert not not_dense_group == [[]]
cluster_list = not_dense_group[0]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 0
assert np.all(cluster.centroid == [(1 / 3), (1 / 3)])
assert cluster.radius == cl1.radius
assert cluster.weight == 3
assert cluster.case_ids == ['0', '1']
with pytest.raises(IndexError):
assert cluster_list[1]
cluster_list = not_dense_group[1]
cluster = cluster_list[0]
assert isinstance(cluster, Cluster)
assert cluster.id == 1
assert np.all(cluster.centroid == [0.5, 0.5])
assert cluster.radius == cl2.radius
assert cluster.weight == 2
assert cluster.case_ids == ['2', '3']
with pytest.raises(IndexError):
assert not_dense_group[2]
assert cluster_list[1]
def test_add_point(self, denstream):
case_list = []
case_1 = Case('1')
case_1.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_1.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_1.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case_1)
case_2 = Case('2')
case_2.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_2.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_list.append(case_2)
graph = nx.DiGraph()
graph = initialize_graph(graph, case_list)
case_3 = Case('3')
case_3.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_3.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_3.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_3.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
trace_distance, time_distance = extract_case_distances(graph, case_3)
case_3.graph_distance = trace_distance
case_3.time_distance = time_distance
micro_cluster = MicroCluster(10, 2, 0, 0.15)
micro_cluster.CF = np.array([0.5, -0.5])
micro_cluster.CF2 = np.array([0.5, -0.1])
micro_cluster.weight = 10
denstream.p_micro_clusters.append(micro_cluster)
micro_cluster = MicroCluster(11, 2, 0, 0.15)
micro_cluster.CF = np.array([0.0, 0.0])
micro_cluster.CF2 = np.array([0.0, 0.0])
micro_cluster.weight = 5
denstream.o_micro_clusters.append(micro_cluster)
denstream.mc_id = 2
mc_id = denstream.add_point(case_3)
assert mc_id == 2
assert len(denstream.o_micro_clusters) == 2
assert denstream.o_micro_clusters[1].radius == 0
assert denstream.o_micro_clusters[1].weight == 1
assert np.all(denstream.o_micro_clusters[1].CF == case_3.point)
assert np.all(denstream.o_micro_clusters[1].CF2 == case_3.point *
case_3.point)
cf = denstream.o_micro_clusters[1].CF.copy()
cf2 = denstream.o_micro_clusters[1].CF2.copy()
mc_id = denstream.add_point(case_3)
assert mc_id == 2
assert len(denstream.o_micro_clusters) == 1
assert len(denstream.p_micro_clusters) == 2
assert denstream.p_micro_clusters[1].weight == 2
assert np.all(denstream.p_micro_clusters[1].CF == cf + case_3.point)
assert np.all(denstream.p_micro_clusters[1].CF2 == cf2 +
case_3.point * case_3.point)
def test_update():
case_list = []
micro_cluster = MicroCluster(0, 2, 0, 0.15)
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 33, 20),
})
case.add_event({
"concept:name": "activityE",
"time:timestamp": datetime(2015, 5, 10, 14, 6, 40),
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 1, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 14, 40, 00),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 15, 5, 00),
})
case_list.append(case)
graph = initialize_graph(case_list)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case.distances = calculate_case_distances(graph, case)
cf = micro_cluster.cf1.copy()
cf2 = micro_cluster.cf2.copy()
weight = micro_cluster.weight
micro_cluster.update(case)
assert np.all(micro_cluster.cf1 == cf + case.point)
assert np.all(micro_cluster.cf2 == cf2 + case.point * case.point)
assert micro_cluster.weight == weight + 1
case = case_list[0]
case.distances = calculate_case_distances(graph, case)
cf = micro_cluster.cf1.copy()
cf2 = micro_cluster.cf2.copy()
weight = micro_cluster.weight
micro_cluster.update(case)
assert np.all(micro_cluster.cf1 == cf + case.point)
assert np.all(micro_cluster.cf2 == cf2 + case.point * case.point)
assert micro_cluster.weight == weight + 1
def test_extract_case_distances():
case_list = []
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 33, 20))
case.set_activity('activityE', datetime(2015, 5, 10, 14, 6, 40))
case_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 1, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 14, 40, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 15, 5, 00))
case_list.append(case)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 13, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 13, 00))
case_list.append(case)
case = Case('4')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case)
case = Case('5')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('6')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 15, 30, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 15, 30, 00))
case_list.append(case)
case = Case('7')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 11, 8, 00, 00))
case_list.append(case)
case = Case('8')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 11, 46, 40))
case_list.append(case)
case = Case('9')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('10')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('11')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
graph = nx.DiGraph()
graph = initialize_graph(graph, case_list)
case = Case('12')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 9, 18, 20))
case.set_activity('activityC', datetime(2015, 5, 10, 10, 8, 20))
trace_distance, time_distance = extract_case_distances(graph, case)
assert trace_distance == pytest.approx(0.05)
assert time_distance == pytest.approx(-1.42, rel=1e-2)
def cases_list(self):
case_1 = Case('1')
case_1.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_1.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_1.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_2 = Case('2')
case_2.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_2.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_3 = Case('3')
case_3.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case_3.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 10))
case_3.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_3.set_activity('activityD', datetime(2015, 5, 10, 8, 00, 30))
return [case_3, case_2, case_1]
def simple_graph_with_attributes(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 10,
"attribute_two": 5,
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 0,
"attribute_two": 7,
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
"attribute_one": 2,
"attribute_two": 4,
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 3,
"attribute_two": 10,
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 6,
"attribute_two": 2,
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 8,
"attribute_two": 4,
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 1,
"attribute_two": 12,
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
"attribute_one": 2,
"attribute_two": 2,
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
"attribute_one": 1,
"attribute_two": 1,
})
case_list.append(case)
graph = initialize_graph(case_list, ["attribute_one", "attribute_two"])
return graph
def complex_graph(self):
case_list = []
case = Case('1')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 8, 33, 20))
case.set_activity('activityE', datetime(2015, 5, 10, 14, 6, 40))
case_list.append(case)
case = Case('2')
case.set_activity('activityA', datetime(2015, 5, 10, 1, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 14, 40, 00))
case.set_activity('activityD', datetime(2015, 5, 10, 15, 5, 00))
case_list.append(case)
case = Case('3')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 13, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 13, 00))
case_list.append(case)
case = Case('4')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 20))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 20))
case_list.append(case)
case = Case('5')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('6')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 15, 30, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 15, 30, 00))
case_list.append(case)
case = Case('7')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 11, 8, 00, 00))
case_list.append(case)
case = Case('8')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 11, 46, 40))
case_list.append(case)
case = Case('9')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('10')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityF', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
case = Case('11')
case.set_activity('activityA', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityB', datetime(2015, 5, 10, 8, 00, 00))
case.set_activity('activityC', datetime(2015, 5, 10, 8, 00, 00))
case_list.append(case)
graph = nx.DiGraph(name='complex_graph')
graph = initialize_graph(graph, case_list)
return graph
def test_release_cases_from_memory(self, process):
case1 = Case("1")
case2 = Case("2")
case3 = Case("3")
case4 = Case("4")
case4.add_event({
"concept:name": "activity1",
"time:timestamp": datetime(2015, 5, 10, 8, 30, 00),
})
case1.add_event({
"concept:name": "activity1",
"time:timestamp": datetime(2015, 5, 10, 9, 00, 00),
})
case3.add_event({
"concept:name": "activity1",
"time:timestamp": datetime(2015, 5, 10, 9, 30, 00),
})
case2.add_event({
"concept:name": "activity1",
"time:timestamp": datetime(2015, 5, 10, 10, 00, 00),
})
process.cases = [case2, case3, case1, case4]
process.release_cases_from_memory()
assert process.cases == []
process.cases = [case2, case3, case1, case4]
process.nyquist = 1
process.release_cases_from_memory()
assert process.cases == [case2]
process.cases = [case2, case3, case1, case4]
process.nyquist = 2
process.release_cases_from_memory()
assert process.cases == [case2, case3]
process.cases = [case2, case3, case1, case4]
process.nyquist = 3
process.release_cases_from_memory()
assert process.cases == [case2, case3, case1]
process.cases = [case2, case3, case1, case4]
process.nyquist = 4
process.release_cases_from_memory()
assert process.cases == [case2, case3, case1, case4]
def test_train(self, denstream: DenStream):
# anomalous_denstream = DenStream(2, 0.15, 0.3, 0.1, 4, 1000)
cases_list = []
case = Case("5")
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
})
case.add_event({
"concept:name": "activityE",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 40),
})
cases_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
})
cases_list.append(case)
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
cases_list.append(case)
case = Case("4")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
cases_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
cases_list.append(case)
pmg = initialize_graph(cases_list)
for case in cases_list:
case.distances = calculate_case_distances(pmg, case)
def test_calculate_actual_distances(self):
case_one = Case("Case 1")
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 8, 30, 30)
})
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 8, 45, 30)
})
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 8, 50, 30)
})
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 8, 53, 30)
})
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 8, 59, 30)
})
case_one.add_event({
"concept:name": "",
"time:timestamp": datetime(2015, 5, 10, 9, 1, 30)
})
expected_distances = [900, 300, 180, 360, 120]
assert calculate_case_time_distances(case_one) == expected_distances
def complex_graph(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 33, 20),
})
case.add_event({
"concept:name": "activityE",
"time:timestamp": datetime(2015, 5, 10, 14, 6, 40),
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 1, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 14, 40, 00),
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 15, 5, 00),
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case_list.append(case)
case = Case("4")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_list.append(case)
case = Case("5")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("6")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case.add_event({
"concept:name": "activityF",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case_list.append(case)
case = Case("7")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityF",
"time:timestamp": datetime(2015, 5, 11, 8, 00, 00),
})
case_list.append(case)
case = Case("8")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityF",
"time:timestamp": datetime(2015, 5, 10, 11, 46, 40),
})
case_list.append(case)
case = Case("9")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityF",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("10")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityF",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("11")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
graph = initialize_graph(case_list)
return graph
class TestCalculateCaseDistances:
@pytest.fixture
def simple_graph_with_attributes(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 10,
"attribute_two": 5,
"color": "red",
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 0,
"attribute_two": 7,
"color": "green",
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
"attribute_one": 2,
"attribute_two": 4,
"color": "green",
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 3,
"attribute_two": 10,
"color": "blue",
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 6,
"attribute_two": 2,
"color": "green",
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 8,
"attribute_two": 4,
"color": "red",
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
"attribute_one": 1,
"attribute_two": 12,
"color": "blue",
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
"attribute_one": 2,
"attribute_two": 2,
"color": "blue",
})
case.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
"attribute_one": 1,
"attribute_two": 1,
"color": "blue",
})
case_list.append(case)
graph = initialize_graph(case_list,
["attribute_one", "attribute_two", "color"])
return graph
def test_calculate_case_distances(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity D",
"time:timestamp": datetime(2015, 5, 10, 8, 33, 20),
})
case.add_event({
"concept:name": "Activity E",
"time:timestamp": datetime(2015, 5, 10, 14, 6, 40),
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 1, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 14, 40, 00),
})
case.add_event({
"concept:name": "Activity D",
"time:timestamp": datetime(2015, 5, 10, 15, 5, 00),
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case_list.append(case)
case = Case("4")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_list.append(case)
case = Case("5")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("6")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case_list.append(case)
case = Case("7")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 11, 8, 00, 00),
})
case_list.append(case)
case = Case("8")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 11, 46, 40),
})
case_list.append(case)
case = Case("9")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("10")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("11")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
graph = initialize_graph(case_list)
case = Case("12")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 9, 18, 20),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 10, 8, 20),
})
distances = calculate_case_distances(graph, case)
assert distances.get("graph") == pytest.approx(0.05)
assert distances.get("time") == pytest.approx(-1.42, rel=1e-2)
def test_calculate_case_distances_with_attributes(
self,
simple_graph_with_attributes: nx.DiGraph,
):
case = Case("4")
case.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
"attribute_one": 9,
"attribute_two": 6,
"color": "red",
})
case.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 9, 18, 20),
"attribute_one": 5,
"attribute_two": 10,
"color": "red",
})
case.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 10, 8, 20),
"attribute_one": 2,
"attribute_two": 2,
"color": "green",
})
distances = calculate_case_distances(
simple_graph_with_attributes,
case,
additional_attributes=["attribute_one", "attribute_two", "color"],
)
assert distances.get("attribute_one") == pytest.approx(0.4121,
rel=1e-2)
assert distances.get("attribute_two") == pytest.approx(0.2651,
rel=1e-2)
assert distances.get("color") == pytest.approx(0.6111, rel=1e-2)
def cases_list(self):
case_1 = Case("1")
case_1.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_1.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case_1.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_2 = Case("2")
case_2.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_2.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case_3 = Case("3")
case_3.add_event({
"concept:name": "activityA",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_3.add_event({
"concept:name": "activityB",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 10),
})
case_3.add_event({
"concept:name": "activityC",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_3.add_event({
"concept:name": "activityD",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 30),
})
return [case_3, case_2, case_1]
def test_calculate_case_distances(self):
case_list = []
case = Case("1")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity D",
"time:timestamp": datetime(2015, 5, 10, 8, 33, 20),
})
case.add_event({
"concept:name": "Activity E",
"time:timestamp": datetime(2015, 5, 10, 14, 6, 40),
})
case_list.append(case)
case = Case("2")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 1, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 14, 40, 00),
})
case.add_event({
"concept:name": "Activity D",
"time:timestamp": datetime(2015, 5, 10, 15, 5, 00),
})
case_list.append(case)
case = Case("3")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 13, 00),
})
case_list.append(case)
case = Case("4")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 20),
})
case_list.append(case)
case = Case("5")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("6")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 15, 30, 00),
})
case_list.append(case)
case = Case("7")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 11, 8, 00, 00),
})
case_list.append(case)
case = Case("8")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 11, 46, 40),
})
case_list.append(case)
case = Case("9")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("10")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity F",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
case = Case("11")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case_list.append(case)
graph = initialize_graph(case_list)
case = Case("12")
case.add_event({
"concept:name": "Activity A",
"time:timestamp": datetime(2015, 5, 10, 8, 00, 00),
})
case.add_event({
"concept:name": "Activity B",
"time:timestamp": datetime(2015, 5, 10, 9, 18, 20),
})
case.add_event({
"concept:name": "Activity C",
"time:timestamp": datetime(2015, 5, 10, 10, 8, 20),
})
distances = calculate_case_distances(graph, case)
assert distances.get("graph") == pytest.approx(0.05)
assert distances.get("time") == pytest.approx(-1.42, rel=1e-2)
