def test_load_distributions():
obj, attributes = general.get_instance_of_ModelContainer()
tstConfig = TstConfig()
temp_folder = tstConfig.get_temperory_file_folder()
#<TO-Do> keeping on hold for now
# Will need to create a sample .joblib file and
# load the 'feature_uniques' and 'feature_summaries'
pass
def test_publish():
pass
def test_dump_reference():
pass
def test_create_merge_request():
pass
def test_bind_model():
pass
def test_dump_model():
pass
def test_load_model():
pass
def test_get_local_path():
pass
def test_get_bucket_path():
pass
def test_dump_distributions():
# This is a to-do and we will see how testing of this happens
# at the moment the method writes a file and we need to work out
# where the test file would be written and how it would be cleaned
obj, attributes = general.get_instance_of_ModelContainer()
pass
def test_get_model():
hmlapp,hmlapp_attributes=general.get_instance_of_HmlApp()
#just instantiated so no models in there
assert len(hmlapp.models.keys())==0
model,model_attr=general.get_instance_of_ModelContainer()
hmlapp.models["test"]=model
assert model==hmlapp.get_model("test")
def test_register_model():
hmlapp,hmlapp_attributes=general.get_instance_of_HmlApp()
model,model_attr=general.get_instance_of_ModelContainer()
hmlapp.register_model("test",model)
assert hmlapp.models["test"]==model
assert "test" in hmlapp.inference.models.keys()
assert model in hmlapp.inference.models.values()
def test_load():
#fil_nam=create_job_lib_file()
fil_nam = create_distributions_file()
obj, attributes = general.get_instance_of_ModelContainer()
# tstConfig=TstConfig()
# temp_folder=tstConfig.get_temperory_file_folder()
#<TO-Do> keeping on hold for now
# Will need to create a sample .joblib file and
#fil_nam=create_job_lib_file()
obj.load(fil_nam)
# if no error is thrown mark as success
#<To-Do> Maybe we can check the content of the loads in a refactoring exercise
assert True
def test_analyze_distributions():
obj, attributes = general.get_instance_of_ModelContainer()
df, numDict, catDict, tarDict = data_frame_utility.get_numerical_categorical_dataframe(
row_count=50)
obj.analyze_distributions(df)
actual_cat_feat_value_dict = obj.feature_uniques
actual_feature_desc = obj.feature_summaries
expected_feature_desc = {
ky: val
for ky in list(numDict.keys()) for val in df[ky].describe().to_dict()
}
#comparing categorical features
assert general.compare_dictionaries_equal(actual_cat_feat_value_dict,
catDict)
#comparing numerical features
assert general.compare_dictionaries_equal(actual_feature_desc,
expected_feature_desc)
def test_build_training_matrix():
obj, attributes = general.get_instance_of_ModelContainer()
df, numDict, catDict, tarDict = data_frame_utility.get_numerical_categorical_dataframe(
row_count=50)
list_of_categorical_columns = attributes["features_categorical"]
expected_unique_value_dict = catDict
obj.build_training_matrix(df)
col_names_removed = list(expected_unique_value_dict.keys())
new_added__categorical_columns = [
f"{col}:{val}" for col in col_names_removed
for val in expected_unique_value_dict[col]
]
numerical_columns = list(numDict.keys())
expected_column_count = len(numerical_columns) + len(
new_added__categorical_columns)
matrix = obj.build_training_matrix(df)
actual_column_count = matrix.shape[1] # first row has the column names
#Checking If The Column Count Matches
assert actual_column_count == expected_column_count
def test_init():
obj, expected_values = general.get_instance_of_ModelContainer()
retBool, retLst = general.check_attributes_in_object(obj, expected_values)
assert retBool