def setUp(self):
set_project_path(testenv.MOCK_PROJECT_PATH)
self.path_to_dataset = testenv.MOCK_DATASET
self.path_to_test_dir = testenv.TEST_PATH
self.aux_project_name = testenv.MOCK_PROJECT_NAME
self.path_to_auxproj = testenv.MOCK_PROJECT_PATH
self.project_default_name = testenv.DEFAULT_PROJECT_NAME
# Generate a project
self.p = Project(path=self.path_to_test_dir,
name=self.aux_project_name)
# Add a dataset
self.ds = Dataset.read_file(path=self.path_to_dataset)
self.ds.save()
# Generate subdataset
self.sbds = SubDataset(self.ds, method="k_fold", by=5)
self.sbds.save()
# set apporach
self.approach = Approach(self.p,
"logistic_regression",
self.sbds,
path=str(Path(testenv.TEST_PATH, "lr")))
shutil.copyfile(testenv.APPROACH_EXAMPLE,
str(self.approach.script_path))
self.approach.save()
# generate runs
import_from("test.lr.logistic_regression",
"LogisticRegressionApproach")().run()
def setUp(self):
set_project_path(testenv.MOCK_PROJECT_PATH)
self.path_to_dataset = testenv.IRIS_DATASET
self.path_to_test_dir = testenv.TEST_PATH
self.aux_project_name = testenv.MOCK_PROJECT_NAME
self.path_to_auxproj = testenv.MOCK_PROJECT_PATH
self.project_default_name = testenv.DEFAULT_PROJECT_NAME
# Generate a project
self.p = Project(path=self.path_to_test_dir,
name=self.aux_project_name)
# Add a dataset
self.ds = Dataset.read_file(path=self.path_to_dataset, )
self.ds.save()
# Generate subdataset
self.sbds = SubDataset(self.ds, method="k_fold", by=5)
self.sbds.save()
# set apporach
self.approach = Approach(self.p,
"decision_tree",
self.sbds,
path=str(Path(testenv.TEST_PATH, "dt")))
shutil.copyfile(testenv.IRIS_APPROACH, str(self.approach.script_path))
self.approach.save()
# generate runs
import_from("test.dt.decision_tree", "DecisionTreeApproach")().run()
def test_generate_subdataset(self, method="train_test", by=0.8):
"""
Creates a Project and a Dataset and links each other. Then creates a SubDataset out of the created Dataset.
The subdataset are created using method and by paramenters
Asserts
-------
- Indicies are not None
"""
Project(path=self.path_to_test_dir, name=self.aux_project_name)
ds = Dataset.read_file(path=self.path_to_dataset)
ds.save()
sbds = SubDataset(dataset=ds, method=method, by=by)
self.assertIsNotNone(sbds.indices)
return sbds
def test_create_dataset(self):
"""
Takes a dataset file and constructs a new dataset
Asserts
-------
- If the object created is an instance of a Dataset object
- If the datasource attribute is an instance of a FileDataset object
Returns
-------
A Dataset object instance created from the dataset file
"""
Project(name=self.aux_project_name, path=self.path_to_test_dir)
ds = Dataset.read_file(path=self.path_to_dataset,
first_line_heading=False)
self.assertIsInstance(ds, Dataset)
self.assertIsInstance(ds.datasource, FileDatasource)
return ds
def test_get_latest_subdataset(self):
p = self.test_load_project()
ds = Dataset.read_file(path=testenv.MOCK_DATASET,
first_line_heading=False)
ds.save()
method = "k_fold"
by = 5
sbds1 = SubDataset(ds, method=method, by=by)
sbds1.save()
sbds2 = SubDataset(ds, method=method, by=by*2)
sbds2.save()
l_sbds = p.get_last_subdataset()
self.assertEqual(l_sbds.id, sbds2.id)
def setUp(self):
set_project_path(testenv.MOCK_PROJECT_PATH)
self.p = Project(path=testenv.TEST_PATH,
name=testenv.MOCK_PROJECT_NAME)
self.ds = Dataset.read_file(path=testenv.MOCK_DATASET,
first_line_heading=False)
self.ds.save()
self.sbds = SubDataset(self.ds, method="k_fold", by=5)
self.sbds.save()
self.approach = Approach(self.p,
"logistic_regression",
self.sbds,
path=str(Path(testenv.TEST_PATH, "lr")))
shutil.copyfile(testenv.APPROACH_EXAMPLE,
str(self.approach.script_path))
self.approach.save()
def setUp(self):
set_project_path(testenv.MOCK_PROJECT_PATH)
self.path_to_dataset = testenv.MOCK_DATASET
self.path_to_test_dir = testenv.TEST_PATH
self.aux_project_name = testenv.MOCK_PROJECT_NAME
self.path_to_auxproj = testenv.MOCK_PROJECT_PATH
self.project_default_name = testenv.DEFAULT_PROJECT_NAME
self.p = Project(path=self.path_to_test_dir,
name=self.aux_project_name)
self.ds = Dataset.read_file(path=self.path_to_dataset)
self.ds.save()
self.sbds = SubDataset(self.ds, method="k_fold", by=5)
self.sbds.save()
self.approach = Approach(self.p, "test_approach", self.sbds)
shutil.copyfile(testenv.APPROACH_EXAMPLE,
str(self.approach.script_path))
self.approach.save()
# create new project
path_to_project = str(Path(".").absolute())
project_name = "test_project"
project_path = Path(path_to_project, project_name)
# If project exist delete and recreate it
if project_path.is_dir():
proj = Project.load(project_path)
clean(proj)
proj = Project(name="test_project", path=path_to_project)
# add a datasource
path_to_dataset = str(Path(r"./test/resources/test_dataset.csv").absolute())
ds = Dataset.read_file(path_to_dataset)
ds.set_project_path(proj.path)
ds.save()
# create subdataset
sbds = SubDataset(ds, method="k_fold", by=5)
sbds.save()
# set apporach
example_approach_path = r"./test/resources/approach_example.py"
param_path = r"./test/resources/parameters_example.yml"
a = Approach(proj, "approach_example", sbds)
shutil.copyfile(example_approach_path, str(a.script_path))
shutil.copyfile(param_path, str(a.params_path))
a.save()
def test_automatically_detect_regression(self):
Project(name=self.aux_project_name, path=self.path_to_test_dir)
ds = Dataset.read_file("test/resources/housing.csv",
label="median_house_value")
self.assertEqual(ds.problem_type, "regression")
def test_automatically_detect_binary_clf(self):
Project(name=self.aux_project_name, path=self.path_to_test_dir)
ds = Dataset.read_file("test/resources/titanic.csv", label="Survived")
self.assertEqual(ds.problem_type, "binary_clf")
def test_automatically_detect_clf(self):
Project(name=self.aux_project_name, path=self.path_to_test_dir)
ds = Dataset.read_file("test/resources/Iris.csv")
self.assertEqual(ds.problem_type, "clf")