def classified_failures(test_job, text_log_errors_failure_lines, test_matcher,
failure_classifications):
from treeherder.model.models import ClassifiedFailure
from treeherder.services.elasticsearch import refresh_index
_, failure_lines = text_log_errors_failure_lines
classified_failures = []
for failure_line in failure_lines:
if failure_line.job_guid == test_job.guid:
classified_failure = ClassifiedFailure()
classified_failure.save()
failure_line.error.set_classification(
test_matcher,
classified_failure,
mark_best=True,
)
classified_failures.append(classified_failure)
if settings.ELASTICSEARCH_URL:
refresh_index()
return classified_failures
def test_cycle_all_but_one_job(test_repository, failure_classifications,
sample_data, sample_push, mock_log_parser,
elasticsearch, failure_lines):
"""
Test cycling all but one job in a group of jobs to confirm there are no
unexpected deletions
"""
job_data = sample_data.job_data[:20]
test_utils.do_job_ingestion(test_repository, job_data, sample_push, False)
# one job should not be deleted, set its submit time to now
job_not_deleted = Job.objects.get(id=2)
job_not_deleted.submit_time = datetime.datetime.now()
job_not_deleted.save()
extra_objects = {
'failure_lines': (FailureLine,
create_failure_lines(job_not_deleted,
[(test_line, {}),
(test_line, {
"subtest": "subtest2"
})])),
'job_details': (JobDetail, [
JobDetail.objects.create(job=job_not_deleted,
title='test',
value='testvalue')
])
}
# set other job's submit time to be a week ago from now
cycle_date_ts = datetime.datetime.now() - datetime.timedelta(weeks=1)
for job in Job.objects.all().exclude(id=job_not_deleted.id):
job.submit_time = cycle_date_ts
job.save()
num_job_logs_to_be_deleted = JobLog.objects.all().exclude(
id=job_not_deleted.id).count()
num_job_logs_before = JobLog.objects.count()
call_command('cycle_data',
'from:treeherder',
sleep_time=0,
days=1,
debug=True)
assert Job.objects.count() == 1
assert JobLog.objects.count() == (num_job_logs_before -
num_job_logs_to_be_deleted)
for (object_type, objects) in extra_objects.values():
assert (set(item.id for item in object_type.objects.all()) == set(
item.id for item in objects))
if settings.ELASTICSEARCH_URL:
refresh_index()
# get all documents
indexed_ids = set(int(item['_id']) for item in all_documents())
expected = set(item.id for item in extra_objects["failure_lines"][1])
assert indexed_ids == expected
def elasticsearch(request):
from treeherder.services.elasticsearch import reinit_index, refresh_index
if not settings.ELASTICSEARCH_URL:
return
reinit_index()
refresh_index()
def elasticsearch(request):
from treeherder.services.elasticsearch import reinit_index, refresh_index
if not settings.ELASTICSEARCH_URL:
return
reinit_index()
refresh_index()
def test_cycle_all_but_one_job(test_repository, failure_classifications, sample_data,
sample_push, mock_log_parser, elasticsearch,
failure_lines):
"""
Test cycling all but one job in a group of jobs to confirm there are no
unexpected deletions
"""
job_data = sample_data.job_data[:20]
test_utils.do_job_ingestion(test_repository, job_data, sample_push, False)
# one job should not be deleted, set its submit time to now
job_not_deleted = Job.objects.get(id=2)
job_not_deleted.submit_time = datetime.datetime.now()
job_not_deleted.save()
extra_objects = {
'failure_lines': (FailureLine,
create_failure_lines(
job_not_deleted,
[(test_line, {}),
(test_line, {"subtest": "subtest2"})])),
'job_details': (JobDetail, [JobDetail.objects.create(
job=job_not_deleted,
title='test',
value='testvalue')])
}
# set other job's submit time to be a week ago from now
cycle_date_ts = datetime.datetime.now() - datetime.timedelta(weeks=1)
for job in Job.objects.all().exclude(id=job_not_deleted.id):
job.submit_time = cycle_date_ts
job.save()
num_job_logs_to_be_deleted = JobLog.objects.all().exclude(
id=job_not_deleted.id).count()
num_job_logs_before = JobLog.objects.count()
call_command('cycle_data', sleep_time=0, days=1, debug=True)
refresh_index()
assert Job.objects.count() == 1
assert JobLog.objects.count() == (num_job_logs_before -
num_job_logs_to_be_deleted)
for (object_type, objects) in extra_objects.values():
assert (set(item.id for item in object_type.objects.all()) ==
set(item.id for item in objects))
# get all documents
indexed_ids = set(int(item['_id']) for item in all_documents())
expected = set(item.id for item in extra_objects["failure_lines"][1])
assert indexed_ids == expected
def test_store_no_subtest(elasticsearch):
doc = {
'job_guid': '1234',
'test': 'test',
'status': 'FAIL',
'expected': 'PASS',
'message': 'Example',
}
es_conn.index(INDEX_NAME, DOC_TYPE, doc)
refresh_index()
docs = list(all_documents())
assert 'subtest' not in docs[0]['_source']
assert count_index() == 1
def test_store_no_subtest(elasticsearch):
doc = {
'job_guid': '1234',
'test': 'test',
'status': 'FAIL',
'expected': 'PASS',
'message': 'Example',
}
es_conn.index(INDEX_NAME, DOC_TYPE, doc)
refresh_index()
docs = list(all_documents())
assert 'subtest' not in docs[0]['_source']
assert count_index() == 1
def create_failure_lines(job, failure_line_list,
start_line=0):
failure_lines = []
for i, (base_data, updates) in enumerate(failure_line_list[start_line:]):
data = {"job_guid": job.guid,
"repository": job.repository,
"line": i + start_line}
data.update(base_data)
data.update(updates)
failure_line = FailureLine(**data)
failure_line.save()
failure_line.elastic_search_insert()
failure_lines.append(failure_line)
if settings.ELASTICSEARCH_URL:
refresh_index()
return failure_lines
def create_failure_lines(job, failure_line_list, start_line=0):
failure_lines = []
for i, (base_data, updates) in enumerate(failure_line_list[start_line:]):
data = {
"job_guid": job.guid,
"repository": job.repository,
"line": i + start_line
}
data.update(base_data)
data.update(updates)
failure_line = FailureLine(**data)
failure_line.save()
failure_line.elastic_search_insert()
failure_lines.append(failure_line)
if settings.ELASTICSEARCH_URL:
refresh_index()
return failure_lines
def classified_failures(test_job, text_log_errors_failure_lines, test_matcher,
failure_classifications):
from treeherder.model.models import ClassifiedFailure
from treeherder.services.elasticsearch import refresh_index
_, failure_lines = text_log_errors_failure_lines
classified_failures = []
for failure_line in failure_lines:
if failure_line.job_guid == test_job.guid:
classified_failure = ClassifiedFailure.objects.create()
failure_line.error.create_match(test_matcher, classified_failure)
mark_best_classification(failure_line.error, classified_failure)
classified_failures.append(classified_failure)
if settings.ELASTICSEARCH_URL:
refresh_index()
return classified_failures
