def test_score_from_params_clustering():
np.random.seed(123)
X = np.random.randn(100, 10)
kmeans = KMeans(n_clusters=3, random_state=123)
result1 = score_from_params_clustering(kmeans, X, scoring=["silhouette", "davies_bouldin"])
with pytest.raises(sklearn.exceptions.NotFittedError):
kmeans.predict(X)
assert isinstance(result1, pd.DataFrame)
assert list(result1.columns) == ["test_silhouette", "test_davies_bouldin", "fit_time", "score_time"]
assert len(result1) == 1
kmeans = KMeans(n_clusters=3, random_state=123)
result2, yhat = score_from_params_clustering(
kmeans, X, scoring=["silhouette", "davies_bouldin"], return_predict=True
)
with pytest.raises(sklearn.exceptions.NotFittedError):
kmeans.predict(X)
assert isinstance(result2, pd.DataFrame)
assert list(result2.columns) == ["test_silhouette", "test_davies_bouldin", "fit_time", "score_time"]
assert len(result2) == 1
assert yhat.shape == (100,)
assert len(np.unique(yhat)) == 3
assert np.abs(result1.iloc[:, 0] - result2.iloc[:, 0]).max() <= 10 ** (-5)
assert np.abs(result1.iloc[:, 1] - result2.iloc[:, 1]).max() <= 10 ** (-5)
def test_score_from_params_clustering_with_scorer_object():
X = np.random.randn(100, 10)
kmeans = KMeans(n_clusters=3, random_state=123)
result1 = score_from_params_clustering(kmeans, X, scoring=SCORERS["silhouette"])
assert result1.shape[0] == 1
assert isinstance(result1, pd.DataFrame)
result2 = score_from_params_clustering(kmeans, X, scoring="silhouette")
assert result2.shape[0] == 1
assert isinstance(result2, pd.DataFrame)
assert np.abs(result1.iloc[:, 0] - result2.iloc[:, 0]).max() <= 10 ** (-5)
result1 = score_from_params_clustering(kmeans, X, scoring=SCORERS["calinski_harabasz"])
assert result1.shape[0] == 1
assert isinstance(result1, pd.DataFrame)
result2 = score_from_params_clustering(kmeans, X, scoring="calinski_harabasz")
assert result2.shape[0] == 1
assert isinstance(result2, pd.DataFrame)
assert np.abs(result1.iloc[:, 0] - result2.iloc[:, 0]).max() <= 10 ** (-5)
result1 = score_from_params_clustering(kmeans, X, scoring=SCORERS["davies_bouldin"])
assert result1.shape[0] == 1
assert isinstance(result1, pd.DataFrame)
result2 = score_from_params_clustering(kmeans, X, scoring="davies_bouldin")
assert result2.shape[0] == 1
assert isinstance(result2, pd.DataFrame)
assert np.abs(result1.iloc[:, 0] - result2.iloc[:, 0]).max() <= 10 ** (-5)
def test_score_from_params(x_data_type, shuffle, graph_pipeline):
np.random.seed(123)
X = np.random.randn(100, 10)
X = convert_generic(X, output_type=x_data_type)
if x_data_type == DataTypes.DataFrame:
X.columns = ["col_%d" % i for i in range(X.shape[1])]
if shuffle:
ii = np.arange(X.shape[0])
np.random.shuffle(ii)
if isinstance(X, pd.DataFrame):
X = X.loc[ii, :]
else:
X = X[ii, :]
scoring = ["silhouette", "davies_bouldin", "calinski_harabasz"]
if graph_pipeline:
estimator = GraphPipeline(
{"pt": DebugPassThrough(), "lg": KMeans(n_clusters=3, random_state=123)}, edges=[("pt", "lg")]
)
else:
estimator = KMeans(n_clusters=3, random_state=123)
##################
### Only score ###
##################
res = score_from_params_clustering(estimator, X, scoring=scoring, verbose=0)
assert isinstance(res, pd.DataFrame)
assert res.shape[0] == 1
for s in scoring:
assert ("test_" + s) in set(res.columns)
with pytest.raises(NotFittedError):
estimator.predict(X)
##########################
### Score + Prediction ###
##########################
res, label = score_from_params_clustering(estimator, X, scoring=scoring, verbose=0, return_predict=True)
assert isinstance(res, pd.DataFrame)
assert res.shape[0] == 1
for s in scoring:
assert ("test_" + s) in set(res.columns)
assert isinstance(label, np.ndarray)
assert len(np.unique(label)) == 3
with pytest.raises(NotFittedError):
estimator.predict(X)
####################
### Predict only ###
####################
res, label = score_from_params_clustering(
estimator, X, scoring=scoring, verbose=0, return_predict=True, no_scoring=True
)
assert len(np.unique(label)) == 3
assert res is None
with pytest.raises(NotFittedError):
estimator.predict(X)