def test_set_cursor(self, base):
model_test = DummyTreeRegressor("tr_cursor_test", cursor=base.cursor)
# TODO: creat a new cursor
model_test.set_cursor(base.cursor)
model_test.drop()
model_test.fit("public.tr_data", ["gender"], "transportation")
base.cursor.execute(
"SELECT model_name FROM models WHERE model_name = 'tr_cursor_test'"
)
assert base.cursor.fetchone()[0] == "tr_cursor_test"
model_test.drop()
def test_contour(self, titanic_vd):
model_test = DummyTreeRegressor("model_contour", )
model_test.drop()
model_test.fit(
titanic_vd,
["age", "fare"],
"survived",
)
result = model_test.contour()
assert len(result.get_default_bbox_extra_artists()) == 34
model_test.drop()
def model(base, tr_data_vd):
base.cursor.execute("DROP MODEL IF EXISTS tr_model_test")
base.cursor.execute(
"SELECT rf_regressor('tr_model_test', 'public.tr_data', 'TransPortation', '*' USING PARAMETERS exclude_columns='id, transportation', mtry=4, ntree=1, max_breadth=1e9, sampling_size=1, max_depth=100, min_leaf_size=1, min_info_gain=0.0, nbins=1000, seed=1, id_column='id')"
)
# I could use load_model but it is buggy
model_class = DummyTreeRegressor("tr_model_test", cursor=base.cursor)
model_class.input_relation = "public.tr_data"
model_class.test_relation = model_class.input_relation
model_class.X = ['"Gender"', '"owned cars"', '"cost"', '"income"']
model_class.y = '"TransPortation"'
yield model_class
model_class.drop()
def test_model_from_vDF(self, base, tr_data_vd):
base.cursor.execute("DROP MODEL IF EXISTS tr_from_vDF")
model_test = DummyTreeRegressor("tr_from_vDF", cursor=base.cursor)
model_test.fit(tr_data_vd, ["gender"], "transportation")
base.cursor.execute(
"SELECT model_name FROM models WHERE model_name = 'tr_from_vDF'")
assert base.cursor.fetchone()[0] == "tr_from_vDF"
model_test.drop()
def test_to_sklearn(self, base):
base.cursor.execute("DROP MODEL IF EXISTS tr_model_sk_test")
base.cursor.execute(
"SELECT rf_regressor('tr_model_sk_test', 'public.tr_data', 'TransPortation', '\"owned cars\"' USING PARAMETERS mtry=1, ntree=1, max_breadth=1e9, sampling_size=1, max_depth=100, min_leaf_size=1, min_info_gain=0.0, nbins=1000, seed=1, id_column='id')"
)
# I could use load_model but it is buggy
model_sk = DummyTreeRegressor("tr_model_sk_test", cursor=base.cursor)
model_sk.input_relation = "public.tr_data"
model_sk.test_relation = model_sk.input_relation
model_sk.X = ['"owned cars"']
model_sk.y = "TransPortation"
md = model_sk.to_sklearn()
model_sk.cursor.execute(
"SELECT PREDICT_RF_REGRESSOR(1 USING PARAMETERS model_name = '{}', match_by_pos=True)"
.format(model_sk.name))
prediction = model_sk.cursor.fetchone()[0]
assert prediction == pytest.approx(md.predict([1])[0])
model_sk.drop()
def test_drop(self):
current_cursor().execute("DROP MODEL IF EXISTS tr_model_test_drop")
model_test = DummyTreeRegressor("tr_model_test_drop", )
model_test.fit(
"public.tr_data",
["Gender", '"owned cars"', "cost", "income"],
"TransPortation",
)
current_cursor().execute(
"SELECT model_name FROM models WHERE model_name = 'tr_model_test_drop'"
)
assert current_cursor().fetchone()[0] == "tr_model_test_drop"
model_test.drop()
current_cursor().execute(
"SELECT model_name FROM models WHERE model_name = 'tr_model_test_drop'"
)
assert current_cursor().fetchone() is None
def test_repr(self, model):
assert "SELECT rf_regressor('public.tr_model_test'," in model.__repr__(
)
model_repr = DummyTreeRegressor("RF_repr")
model_repr.drop()
assert model_repr.__repr__() == "<RandomForestRegressor>"