-
Notifications
You must be signed in to change notification settings - Fork 370
/
runner.py
219 lines (192 loc) · 7.88 KB
/
runner.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
import os
import re
import sys
import argparse
import time
def get_first_level_path(path):
dir_name = os.path.dirname(path)
parts = dir_name.split('/')
if len(parts) == 2: # <package-name>/test
return path # first level *.py
return "/".join(parts[:3]) # <package-name>/test/<dir>
def run(pytest, output_file, repeatUntilFailure, testDir, test_slice):
if repeatUntilFailure:
log("'repeatUntilFailure' is set")
log("Is going to repeat the test suite until failure")
log("Preparing test suite with {}".format(pytest))
testListFile = "test_list.txt"
collect_status = os.system('python3 -m pytest -q --collect-only {} > {}'.format(testDir, testListFile))
if collect_status != 0:
log("Test suit preparation error {}".format(collect_status))
return -1
log("Reading collected modules file")
collectedData = open(testListFile).read()
os.remove(testListFile)
log("Collecting modules")
# check errors during collect
if re.findall("={5,} ERRORS ={5,}", collectedData):
log("Errors found during collection")
log(collectedData)
return -1
testList = re.findall('({}.*?)::'.format(testDir), collectedData)
# testList = list(set(os.path.dirname(t) for t in testList))
first_level_tests = set()
for path in testList:
p = get_first_level_path(path)
first_level_tests.add(p)
first_level_tests = sorted(list(first_level_tests))
test_list_sliced = first_level_tests[test_slice[0] - 1::test_slice[1]]
log("Found {} tests in {} modules, sliced {} test modules".
format(len(testList), len(first_level_tests), len(test_list_sliced)))
if not testList:
m = re.search("errors during collection", collectedData)
if m:
log(collectedData)
return -1
retVal = 0
totalPassed = 0
totalFailed = 0
totalSkipped = 0
totalErros = 0
runsCount = 0
allFailedTests = []
allErrorTests = []
failureData = []
testRep = 'currentTestReport-{}-{}-{}.txt'.format(testDir, output_file, time.time())
passPat = re.compile(".* ([0-9]+) passed.*$")
skipPat = re.compile(".* ([0-9]+) skipped.*$")
failPat = re.compile(".* ([0-9]+) failed.*$")
errPat = re.compile(".* ([0-9]+) error.*$")
failedTestPat = re.compile('____ (test.+) ____')
errorTestPat = re.compile('____ (ERROR.+) ____')
while True:
for i, tests in enumerate(test_list_sliced):
testResults = '--junitxml={}-test-results.xml'.format(tests.split("/")[-1])
testStartTime = time.time()
pytest_cmd = '{} {} {} > {}'.format(pytest, testResults, tests, testRep)
log(pytest_cmd)
r = os.system(pytest_cmd)
testExecutionTime = time.time() - testStartTime
reportLines = open(testRep).readlines()
output = ''.join(reportLines)
pas = passPat.search(output)
passed = int(pas.groups()[0]) if pas else 0
skp = skipPat.search(output)
skipped = int(skp.groups()[0]) if skp else 0
if r:
fai = failPat.search(output)
err = errPat.search(output)
if not (fai or err):
log("Non zero return value from {} run but no failures "
"or errors reported".format(tests))
log(output)
return -1
failed = int(fai.groups()[0]) if fai else 0
errors = int(err.groups()[0]) if err else 0
failedNames = []
errorNames = []
startedF = None
startedE = None
for line in reportLines:
if '= FAILURES =' in line:
startedF = True
startedE = None
continue
if '= ERRORS =' in line:
startedF = None
startedE = True
continue
if startedF:
failureData.append(line)
m = failedTestPat.search(line)
if m:
failedNames.append(m.groups()[0])
if startedE:
failureData.append(line)
m = errorTestPat.search(line)
if m:
errorNames.append(m.groups()[0])
else:
failed = 0
errors = 0
log('\tIn {}, {} passed, {} failed, {} errors, {} skipped, {:.1f}s time '
'({}/{} progress)'.
format(tests, passed, errors, failed, skipped,
testExecutionTime, i + 1, len(test_list_sliced)))
if failed:
logError("Failed tests: {}".format(', '.join(failedNames)))
for nm in failedNames:
allFailedTests.append((tests, nm))
if errors:
logError("Error in tests: {}".format(', '.join(errorNames)))
for nm in errorNames:
allErrorTests.append((tests, nm))
retVal += r
totalPassed += passed
totalFailed += failed
totalErros += errors
totalSkipped += skipped
runsCount += 1
if repeatUntilFailure:
if totalFailed or totalErros:
break # repeatUntilFailure set and failures happened
else:
logSuccess('Run #{} was successful'.format(runsCount))
log('\n\n')
else:
break # just one run
summaryMsg = 'Total {} runs {} passed, {} failed, {} errors, {} skipped'.\
format(runsCount, totalPassed, totalFailed, totalErros, totalSkipped)
log(summaryMsg)
if totalFailed:
log("Failed tests:")
for fm, fn in allFailedTests:
log('{}:{}'.format(fm, fn))
if totalErros:
log("Error in tests:")
for fm, fn in allErrorTests:
log('{}:{}'.format(fm, fn))
if failureData and output_file:
log("Writing failure data in Test-Report.txt")
with open(output_file, 'w') as f:
f.write(summaryMsg)
f.write(''.join(failureData))
if os.path.exists(testRep):
os.remove(testRep)
log("Tests run. Returning {}".format(retVal))
return retVal
def log(msg):
return print(msg, flush=True)
def logError(msg):
return print('\x1b[0;30;41m' + msg + '\x1b[0m', flush=True)
def logSuccess(msg):
return print('\x1b[6;30;42m' + msg + '\x1b[0m')
def parse_test_slice(x):
inc, step = x.split('/')[:2]
return int(inc), int(step)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--pytest', type=str,
help='pytest instance', default='python -m pytest')
parser.add_argument('--output', type=str,
help='result file', default='../Test-Report.txt')
parser.add_argument('--dir', type=str,
help='dir with test', default='.')
parser.add_argument('--nooutput',
help='no result file', action="store_true")
parser.add_argument('--repeat', dest='repeatUntilFailure',
action="store_true",
help='repeat the test suite until failure')
parser.add_argument('--test-only-slice', dest='test_slice', type=parse_test_slice,
default=(1, 1),
help='example: `1/3` -- run only first third part of the tests, \
it is needed only for parallel testing')
args = parser.parse_args()
r = run(
pytest=args.pytest,
output_file=args.output if not args.nooutput else None,
repeatUntilFailure=args.repeatUntilFailure,
testDir=args.dir,
test_slice=args.test_slice
)
sys.exit(0 if r == 0 else 1)