-
Notifications
You must be signed in to change notification settings - Fork 1
/
evaluation.py
314 lines (248 loc) · 10.2 KB
/
evaluation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
class EvaluationMetric(object):
"""
Evaluation Metric
"""
def __init__(self):
self.available_time = 0
self.reset()
def update_available_time(self, available_time):
self.available_time = available_time
def reset(self):
self.scheduled_testcases = []
self.unscheduled_testcases = []
self.detection_ranks = []
self.detection_ranks_time = []
self.detection_ranks_failures = []
# Time to Fail (rank value)
self.ttf = 0
self.ttf_duration = 0
# APFD or NAPFD value
self.fitness = 0
self.detected_failures = 0
self.undetected_failures = 0
self.recall = 0
self.avg_precision = 0
# APFDc (to compute at same time, for instance, with NAPFD)
self.cost = 0
def evaluate(self, test_suite):
raise NotImplementedError('This method must be override')
class NAPFDMetric(EvaluationMetric):
"""
Normalized Average Percentage of Faults Detected (NAPFD) Metric based
"""
def __init__(self):
super().__init__()
def __str__(self):
return 'NAPFD'
def evaluate(self, test_suite):
super().reset()
rank_counter = 1
total_failure_count = 0
scheduled_time = 0
costs = []
failures = 0
self.detected_failures = 0
# We consider the faults are different, that is, a fault is only revelead by only a test case
# Build prefix sum of durations to find cut off point
for row in test_suite:
total_failure_count += row['NumErrors']
failures += row['Verdict']
costs.append(row['Duration'])
# Time spent to fail
if len(self.detection_ranks_time) == 0:
self.ttf_duration += row['Duration']
if scheduled_time + row['Duration'] <= self.available_time:
# If the Verdict is "Failed"
if row['NumErrors'] > 0:
self.detected_failures += row['NumErrors'] * rank_counter
self.detection_ranks.append(rank_counter)
# Individual information
self.detection_ranks_failures.append(row['NumErrors'])
self.detection_ranks_time.append(row['Duration'])
scheduled_time += row['Duration']
self.scheduled_testcases.append(row['Name'])
rank_counter += 1
else:
self.unscheduled_testcases.append(row['Name'])
self.undetected_failures += row['NumErrors']
if total_failure_count > 0:
# Time to Fail (rank value)
self.ttf = self.detection_ranks[0] if self.detection_ranks else 0
self.recall = sum(self.detection_ranks_failures) / \
total_failure_count
self.avg_precision = 123
p = self.recall if self.undetected_failures > 0 else 1
no_testcases = len(test_suite)
# NAPFD
self.fitness = p - self.detected_failures / \
(total_failure_count * no_testcases) + p / (2 * no_testcases)
# APFDc
self.cost = sum([sum(costs[i - 1:]) - 0.5 * costs[i - 1] for i in self.detection_ranks]) / (
sum(costs) * failures)
else:
# Time to Fail (rank value)
self.ttf = -1
self.recall = 1
self.avg_precision = 1
# NAPFD
self.fitness = 1
# APFDc
self.cost = 1
class NAPFDVerdictMetric(EvaluationMetric):
"""
Normalized Average Percentage of Faults Detected (NAPFD) Metric based on Verdict
"""
def __init__(self):
super().__init__()
def __str__(self):
return 'NAPFDVerdict'
def evaluate(self, test_suite):
super().reset()
rank_counter = 1
total_failure_count = 0
scheduled_time = 0
costs = []
# Build prefix sum of durations to find cut off point
for row in test_suite:
total_failure_count += row['Verdict']
costs.append(row['Duration'])
# Time spent to fail
if len(self.detection_ranks_time) == 0:
self.ttf_duration += row['Duration']
if scheduled_time + row['Duration'] <= self.available_time:
# If the Verdict is "Failed"
if row['Verdict']:
self.detection_ranks.append(rank_counter)
# Individual information
self.detection_ranks_failures.append(row['Verdict'])
self.detection_ranks_time.append(row['Duration'])
scheduled_time += row['Duration']
self.scheduled_testcases.append(row['Name'])
rank_counter += 1
else:
self.unscheduled_testcases.append(row['Name'])
self.undetected_failures += row['Verdict']
self.detected_failures = len(self.detection_ranks)
assert self.undetected_failures + self.detected_failures == total_failure_count
if total_failure_count > 0:
# Time to Fail (rank value)
self.ttf = self.detection_ranks[0] if self.detection_ranks else 0
self.recall = self.detected_failures / total_failure_count
self.avg_precision = 123
p = self.recall if self.undetected_failures > 0 else 1
no_testcases = len(test_suite)
# NAPFD
self.fitness = p - sum(self.detection_ranks) / \
(total_failure_count * no_testcases) + p / (2 * no_testcases)
# APFDc
self.cost = sum([sum(costs[i - 1:]) - 0.5 * costs[i - 1] for i in self.detection_ranks]) / (
sum(costs) * total_failure_count)
else:
# Time to Fail (rank value)
self.ttf = -1
self.recall = 1
self.avg_precision = 1
# NAPFD
self.fitness = 1
# APFDc
self.cost = 1
class APFDMetric(EvaluationMetric):
"""
Average Percentage of Faults Detected (APFD) Metric based on Verdict
"""
def __init__(self):
super().__init__()
def __str__(self):
return 'Average Percentage of Faults Detected (APFD)'
def evaluate(self, test_suite):
super().reset()
rank_counter = 1
scheduled_time = 0
total_failure_count = 0
scheduled_time = 0
# Build prefix sum of durations to find cut off point
for row in test_suite:
total_failure_count += row['Verdict']
if scheduled_time + row['Duration'] <= self.available_time:
# If the Verdict is "Failed"
if row['Verdict']:
self.detection_ranks.append(rank_counter)
scheduled_time += row['Duration']
self.scheduled_testcases.append(row['Name'])
rank_counter += 1
else:
self.undetected_failures += row['Verdict']
self.detected_failures = len(self.detection_ranks)
assert self.undetected_failures + self.detected_failures == total_failure_count
if total_failure_count > 0:
# Time to Fail (rank value)
self.ttf = self.detection_ranks[0] if self.detection_ranks else 0
no_testcases = len(test_suite)
# APFD
self.fitness = 1 - sum(self.detection_ranks) / (no_testcases *
total_failure_count) + 1 / (2 * no_testcases)
self.recall = self.detected_failures / total_failure_count
self.avg_precision = 123
else:
# Time to Fail (rank value)
self.ttf = 0
# APFD
self.fitness = 1
self.recall = 1
self.avg_precision = 1
class APFDcMetric(EvaluationMetric):
"""
Average Percentage of Faults Detected with cost (APFDc)
"""
def __init__(self):
super().__init__()
def __str__(self):
return 'Average Percentage of Faults Detected with cost (APFDc)'
def evaluate(self, test_suite):
super().reset()
# The severity for each fault is the same for all
rank_counter = 1
total_failure_count = 0
costs = []
scheduled_time = 0
# Build prefix sum of durations to find cut off point
for row in test_suite:
total_failure_count += row['Verdict']
costs.append(row['Duration'])
# Time spent to fail
if len(self.detection_ranks_time) == 0:
self.ttf_duration += row['Duration']
if scheduled_time + row['Duration'] <= self.available_time:
# If the Verdict is "Failed"
if row['Verdict']:
self.detection_ranks.append(rank_counter)
# Individual information
self.detection_ranks_failures.append(row['Verdict'])
self.detection_ranks_time.append(row['Duration'])
scheduled_time += row['Duration']
self.scheduled_testcases.append(row['Name'])
rank_counter += 1
else:
self.unscheduled_testcases.append(row['Name'])
self.undetected_failures += row['Verdict']
self.detected_failures = len(self.detection_ranks)
assert self.undetected_failures + self.detected_failures == total_failure_count
if total_failure_count > 0:
# Time to Fail (rank value)
self.ttf = self.detection_ranks[0] if self.detection_ranks else 0
detection_cost = 0
for i in self.detection_ranks:
detection_cost += sum(costs[i - 1:]) - 0.5 * costs[i - 1]
no_testcases = len(test_suite)
# APFDc
self.fitness = (detection_cost) / \
(sum(costs) * total_failure_count)
self.recall = self.detected_failures / total_failure_count
self.avg_precision = 123
else:
# Time to Fail (rank value)
self.ttf = -1
# APFDc
self.fitness = 1
self.recall = 1
self.avg_precision = 1