def test_type_check_fail(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
project.add_column(TestCol())
data.integer_col = data.integer_col.astype(float)
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
with pytest.raises(InvalidType):
ds.run()
def test_type_check(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
assert 1
def test_project_get_table(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert hist_data.shape[0] > 0
def test_null_check_fail(data, project):
class TestCol(Column):
column_name = "float_col"
column_type = FloatType()
null = False
force_null = True
project.add_column(TestCol())
data.loc[0, "float_col"] = np.NaN
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
with pytest.raises(NullableError):
ds.run()
def test_multiple_columns(data, project):
class TestCol(Column):
column_name = ["integer_col", "float_col"]
column_type = NumericType()
null = False
force_null = True
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert "integer_col" in hist_data.column_name.values
assert "float_col" in hist_data.column_name.values
def test_function_call_without_arguments_dict_style(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
null = False
force_null = True
functions = [{"function": mean}]
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert hist_data[hist_data.metric == "mean"].value.astype(
float).values[0] == 2.0
def test_other_column(data, project):
class TestCol(Column):
column_name = ["integer_col"]
column_type = IntType()
null = False
force_null = True
functions = [{
"function": correlation_two_columns,
"parameters": {
"column_two": "float_col"
},
}]
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
assert 1
def test_default_checks_done(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert (hist_data[hist_data.metric == "perc_missing"].value.astype(
float).values[0] == 0)
assert (hist_data[(hist_data.metric == "count") & (
hist_data.column_name == "rows")].value.astype(int).values[0] == 3)
assert (hist_data[(hist_data.metric == "count") & (
hist_data.column_name == "columns")].value.astype(int).values[0] == 3)
assert hist_data[hist_data.metric == "dtype"].value.values[0] == "int64"
def test_table_extra_functions(data, project):
@function(return_format=float)
def test_fun(data, column):
return 0
class Table(PandasTable):
columns = ["integer_col"]
infer_schema = False
table_name = "test"
types = {"integer_col": IntType()}
extra_functions = {"integer_col": [{"function": test_fun}]}
project.add_table(Table())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert "test_fun" in hist_data.metric.values
def test_function_call_with_arguments(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
null = False
force_null = True
functions = [{
"function": number_of_outliers,
"parameters": {
"std_away": 2
}
}]
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert (hist_data[hist_data.metric == "number_of_outliers"].value.astype(
int).values[0] == 0)
def test_project_file(data, project):
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
with open(os.path.join(project.config_dir, "projects.json"), "r") as f:
project_file = json.loads(f.read())
assert "test" in project_file
project.delete_from_project_config()
with open(os.path.join(project.config_dir, "projects.json"), "r") as f:
project_file = json.loads(f.read())
assert "test" not in project_file
project.add_to_project_list(schema={
"integer_col": {
"unique": False,
"dtype": "int32",
"nullable": True
}
})
with open(os.path.join(project.config_dir, "projects.json"), "r") as f:
project_file = json.loads(f.read())
assert "test" in project_file
def test_table_equals_columns(data, project):
class Table(PandasTable):
columns = ["integer_col"]
infer_schema = False
table_name = "test"
types = {"integer_col": IntType()}
project.add_table(Table())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert hist_data.shape[0] == 7
project.columns = {}
project.delete_from_project_config()
project.delete_data()
class TestCol(Column):
column_name = "integer_col"
column_type = IntType()
force_type = True
functions = [mean, std, percentage_missing]
project.add_column(TestCol())
ds = DataSet(project=project, batch_name="test")
ds.set_dataset(data)
ds.run()
hist_data = project.get_project_table()
assert hist_data.shape[0] == 7
hist_data = project.get_project_table()
assert hist_data.shape[0] > 0