def test_full_data_check(self):
print("Boost model folder: ", self.boost_model_folder)
decision_maker = BoostingDecisionMaker(self.boost_model_folder)
boost_model_results = self.get_fixture(self.boost_model_results)
tests = [
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.one_hit_search_rs_explained))],
"config":
TestBoostingModel.get_default_config(),
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.two_hits_search_rs_explained))],
"config":
TestBoostingModel.get_default_config(),
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.two_hits_search_rs_explained)),
(self.get_fixture(self.log_message),
self.get_fixture(self.one_hit_search_rs_explained))],
"config":
TestBoostingModel.get_default_config(),
},
{
"elastic_results":
[(self.get_fixture(self.log_message_only_small_logs),
self.get_fixture(self.two_hits_search_rs_small_logs))],
"config":
TestBoostingModel.get_default_config(),
},
]
for idx, test in enumerate(tests):
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"], test["config"],
decision_maker.get_feature_ids())
with sure.ensure('Error in the test case number: {0}', idx):
gathered_data, issue_type_names = _boosting_featurizer.gather_features_info(
)
gathered_data.should.equal(boost_model_results[str(idx)][0],
epsilon=self.epsilon)
predict_label, predict_probability = decision_maker.predict(
gathered_data)
predict_label.tolist().should.equal(
boost_model_results[str(idx)][1], epsilon=self.epsilon)
predict_probability.tolist().should.equal(
boost_model_results[str(idx)][2], epsilon=self.epsilon)
def test_normalize_results(self):
tests = [
{
"elastic_results": [],
"config": TestBoostingFeaturizer.get_default_config(),
"result": [],
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.one_hit_search_rs_explained))],
"config":
TestBoostingFeaturizer.get_default_config(),
"result": [[{
"_score": 158.08437,
"normalized_score": 1.0,
}]],
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.two_hits_search_rs_explained))],
"config":
TestBoostingFeaturizer.get_default_config(),
"result": [[
{
"_score": 158.08437,
"normalized_score": 1.0,
},
{
"_score": 77.53298,
"normalized_score": 0.4904,
},
]],
},
]
for idx, test in enumerate(tests):
with sure.ensure('Error in the test case number: {0}', idx):
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"], test["config"], [])
_boosting_featurizer.all_results.should.have.length_of(
len(test["result"]))
for i in range(len(test["result"])):
for j in range(len(test["result"][i])):
for field in test["result"][i][j]:
elastic_res = _boosting_featurizer.all_results[i][
1][j]
elastic_res[field].should.equal(
test["result"][i][j][field],
epsilon=self.epsilon)
def test_normalize_results(self):
tests = [
{
"elastic_results": [],
"config": TestBoostingFeaturizer.get_default_config(),
"result": [],
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(),
"result": [[{"_score": 158.08437,
"normalized_score": 1.0, }]],
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.two_hits_search_rs_explained, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(),
"result": [[{"_score": 158.08437,
"normalized_score": 1.0,
},
{"_score": 77.53298,
"normalized_score": 0.4904,
}, ]],
},
]
weight_log_sim = weighted_similarity_calculator.\
WeightedSimilarityCalculator(folder=self.weights_folder)
for idx, test in enumerate(tests):
with sure.ensure('Error in the test case number: {0}', idx):
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"],
test["config"],
[],
weighted_log_similarity_calculator=weight_log_sim)
_boosting_featurizer.all_results.should.have.length_of(len(test["result"]))
for i in range(len(test["result"])):
for j in range(len(test["result"][i])):
for field in test["result"][i][j]:
elastic_res = _boosting_featurizer.all_results[i][1][j]
elastic_res[field].should.equal(test["result"][i][j][field],
epsilon=self.epsilon)
def test_find_most_relevant_by_type(self):
tests = [
{
"elastic_results": [],
"config": TestBoostingFeaturizer.get_default_config(),
"result": {},
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.one_hit_search_rs_explained))],
"config":
TestBoostingFeaturizer.get_default_config(),
"result": {
"AB001": {
"mrHit": {
"_score": 158.08437,
"_id": "1"
},
"compared_log": self.get_fixture(self.log_message),
"score": 1.0,
},
}
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.two_hits_search_rs_explained))],
"config":
TestBoostingFeaturizer.get_default_config(),
"result": {
"AB001": {
"mrHit": {
"_score": 158.08437,
"_id": "1"
},
"compared_log": self.get_fixture(self.log_message),
"score": 0.6709,
},
"PB001": {
"mrHit": {
"_score": 77.53298,
"_id": "2"
},
"compared_log": self.get_fixture(self.log_message),
"score": 0.3291,
},
}
},
{
"elastic_results":
[(self.get_fixture(self.log_message),
self.get_fixture(self.two_hits_search_rs_explained)),
(self.get_fixture(self.log_message),
self.get_fixture(self.one_hit_search_rs_explained))],
"config":
TestBoostingFeaturizer.get_default_config(),
"result": {
"AB001": {
"mrHit": {
"_score": 158.08437,
"_id": "1"
},
"compared_log": self.get_fixture(self.log_message),
"score": 0.8031,
},
"PB001": {
"mrHit": {
"_score": 77.53298,
"_id": "2"
},
"compared_log": self.get_fixture(self.log_message),
"score": 0.1969,
},
}
},
]
for idx, test in enumerate(tests):
with sure.ensure('Error in the test case number: {0}', idx):
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"], test["config"], [])
scores_by_issue_type = _boosting_featurizer.find_most_relevant_by_type(
)
scores_by_issue_type.should.have.length_of(len(test["result"]))
for issue_type in test["result"]:
scores_by_issue_type.keys().should.contain(issue_type)
elastic_res = scores_by_issue_type[issue_type]
for field in test["result"][issue_type]:
if type(test["result"][issue_type][field]) != dict:
elastic_res[field].should.equal(
test["result"][issue_type][field],
epsilon=self.epsilon)
else:
for field_dict in test["result"][issue_type][
field]:
result_field_dict = test["result"][issue_type][
field][field_dict]
elastic_res[field][field_dict].should.equal(
result_field_dict, epsilon=self.epsilon)
def test_full_data_check(self):
print("Boost model folder : ", self.boost_model_folder)
print("Weights model folder : ", self.weights_folder)
print("Global defect type model folder : ",
self.global_defect_type_model_folder)
decision_maker = BoostingDecisionMaker(folder=self.boost_model_folder)
boost_model_results = utils.get_fixture(self.boost_model_results,
to_json=True)
tests = []
for log_lines, filter_fields, _decision_maker in [
(-1, ["detected_message", "stacktrace"], decision_maker),
(2, ["message"], decision_maker)
]:
tests.extend([
{
"elastic_results":
[(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained,
to_json=True))],
"config":
self.get_default_config(number_of_log_lines=log_lines,
filter_fields=filter_fields),
"decision_maker":
_decision_maker
},
{
"elastic_results":
[(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.two_hits_search_rs_explained,
to_json=True))],
"config":
self.get_default_config(number_of_log_lines=log_lines,
filter_fields=filter_fields),
"decision_maker":
_decision_maker
},
{
"elastic_results":
[(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.two_hits_search_rs_explained,
to_json=True)),
(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained,
to_json=True))],
"config":
self.get_default_config(number_of_log_lines=log_lines,
filter_fields=filter_fields),
"decision_maker":
_decision_maker
},
{
"elastic_results":
[(utils.get_fixture(self.log_message_only_small_logs,
to_json=True),
utils.get_fixture(self.two_hits_search_rs_small_logs,
to_json=True))],
"config":
self.get_default_config(number_of_log_lines=log_lines,
filter_fields=filter_fields),
"decision_maker":
_decision_maker
},
])
for idx, test in enumerate(tests):
feature_ids = test["decision_maker"].get_feature_ids()
weight_log_sim = None
if self.weights_folder.strip():
weight_log_sim = weighted_similarity_calculator.\
WeightedSimilarityCalculator(folder=self.weights_folder)
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"],
test["config"],
feature_ids,
weighted_log_similarity_calculator=weight_log_sim)
if self.global_defect_type_model_folder.strip():
_boosting_featurizer.set_defect_type_model(
defect_type_model.DefectTypeModel(
folder=self.global_defect_type_model_folder))
with sure.ensure('Error in the test case number: {0}', idx):
gathered_data, issue_type_names = _boosting_featurizer.gather_features_info(
)
gathered_data.should.equal(boost_model_results[str(idx)][0],
epsilon=self.epsilon)
predict_label, predict_probability = test[
"decision_maker"].predict(gathered_data)
predict_label.tolist().should.equal(
boost_model_results[str(idx)][1], epsilon=self.epsilon)
predict_probability.tolist().should.equal(
boost_model_results[str(idx)][2], epsilon=self.epsilon)
def test_filter_by_min_should_match(self):
tests = [
{
"elastic_results": [],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[]),
"result": [],
},
{
"elastic_results": [],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"detected_message", "stacktrace"]),
"result": [],
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"detected_message", "stacktrace"]),
"result": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True))],
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"message"]),
"result": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True))],
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True)),
(utils.get_fixture(self.log_message_wo_stacktrace, to_json=True),
utils.get_fixture(
self.one_hit_search_rs_explained_wo_stacktrace, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"message"]),
"result": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True)),
(utils.get_fixture(self.log_message_wo_stacktrace, to_json=True),
utils.get_fixture(
self.one_hit_search_rs_explained_wo_stacktrace, to_json=True))]
},
{
"elastic_results": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True)),
(utils.get_fixture(self.log_message_wo_stacktrace, to_json=True),
utils.get_fixture(
self.one_hit_search_rs_explained_wo_stacktrace, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"detected_message", "stacktrace"]),
"result": [(utils.get_fixture(self.log_message, to_json=True),
utils.get_fixture(self.one_hit_search_rs_explained, to_json=True)),
(utils.get_fixture(self.log_message_wo_stacktrace, to_json=True),
utils.get_fixture(
self.one_hit_search_rs_explained_wo_stacktrace, to_json=True))]
},
{
"elastic_results": [(utils.get_fixture(self.log_message_only_small_logs, to_json=True),
utils.get_fixture(self.one_hit_search_rs_small_logs, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"detected_message", "stacktrace"]),
"result": []
},
{
"elastic_results": [(utils.get_fixture(self.log_message_only_small_logs, to_json=True),
utils.get_fixture(self.two_hits_search_rs_small_logs, to_json=True))],
"config": TestBoostingFeaturizer.get_default_config(filter_fields=[
"detected_message", "stacktrace"]),
"result": [(utils.get_fixture(self.log_message_only_small_logs, to_json=True),
utils.get_fixture(self.two_hits_search_rs_small_logs, to_json=True))]
},
]
weight_log_sim = weighted_similarity_calculator.\
WeightedSimilarityCalculator(folder=self.weights_folder)
for idx, test in enumerate(tests):
with sure.ensure('Error in the test case number: {0}', idx):
_boosting_featurizer = BoostingFeaturizer(
test["elastic_results"],
test["config"],
[],
weighted_log_similarity_calculator=weight_log_sim)
all_results = test["elastic_results"]
for field in test["config"]["filter_min_should_match"]:
all_results = _boosting_featurizer.filter_by_min_should_match(all_results, field=field)
all_results.should.have.length_of(len(test["result"]))
for idx, (log, hits) in enumerate(all_results):
log["_id"].should.equal(test["result"][idx][0]["_id"])
for i, hit in enumerate(hits["hits"]["hits"]):
hit["_id"].should.equal(test["result"][idx][1]["hits"]["hits"][i]["_id"])