Use ancest-order to process local discretization

Fix local discretization
Refactor tests
Unifiy iris dataset from sklearn with iris.arff

bayesclass/clfs.py

@@ -47,6 +47,12 @@ class BayesBase(BaseEstimator, ClassifierMixin):
     def default_class_name():
         return "class"
 
+    def build_dataset(self):
+        self.dataset_ = pd.DataFrame(
+            self.X_, columns=self.feature_names_in_, dtype=np.int32
+        )
+        self.dataset_[self.class_name_] = self.y_
+
     def _check_params_fit(self, X, y, expected_args, kwargs):
         """Check the common parameters passed to fit"""
         # Check that X and y have correct shape
@@ -64,6 +70,10 @@ class BayesBase(BaseEstimator, ClassifierMixin):
             else:
                 raise ValueError(f"Unexpected argument: {key}")
         self.feature_names_in_ = self.features_
+        # used for local discretization
+        self.indexed_features_ = {
+            feature: i for i, feature in enumerate(self.features_)
+        }
         if self.random_state is not None:
             random.seed(self.random_state)
         if len(self.feature_names_in_) != X.shape[1]:
@@ -125,10 +135,7 @@ class BayesBase(BaseEstimator, ClassifierMixin):
         # Store the information needed to build the model
         self.X_ = X_
         self.y_ = y_
-        self.dataset_ = pd.DataFrame(
-            self.X_, columns=self.feature_names_in_, dtype=np.int32
-        )
-        self.dataset_[self.class_name_] = self.y_
+        self.build_dataset()
         # Build the DAG
         self._build()
         # Train the model
@@ -660,14 +667,8 @@ class Proposal(BaseEstimator):
         # Build the model
         super(self.class_type, self.estimator).fit(self.Xd, y, **kwargs)
         # Local discretization based on the model
-        features = kwargs["features"]
-        # assign indices to feature names
-        self.idx_features_ = dict(list(zip(features, range(len(features)))))
-        upgraded, self.Xd = self._local_discretization()
+        self._local_discretization()
         # self.check_integrity("fit", self.Xd)
-        if upgraded:
-            kwargs = self.update_kwargs(y, kwargs)
-            super(self.class_type, self.estimator).fit(self.Xd, y, **kwargs)
         self.fitted_ = True
         return self
 
@@ -705,27 +706,45 @@ class Proposal(BaseEstimator):
 
     def _local_discretization(self):
         """Discretize each feature with its fathers and the class"""
-        res = self.Xd.copy()
-        upgraded = False
-        # print("-" * 80)
-        for idx, feature in enumerate(self.estimator.feature_names_in_):
+        upgrade = False
+        # order of local discretization is important. no good 0, 1, 2...
+        ancestral_order = list(nx.topological_sort(self.estimator.dag_))
+        for feature in ancestral_order:
+            if feature == self.estimator.class_name_:
+                continue
+            idx = self.estimator.indexed_features_[feature]
             fathers = self.estimator.dag_.get_parents(feature)
             if len(fathers) > 1:
-                # print(
-                #     "Discretizing " + feature + " with " + str(fathers),
-                #     end=" ",
-                # )
                 # First remove the class name as it will be added later
                 fathers.remove(self.estimator.class_name_)
                 # Get the fathers indices
-                features = [self.idx_features_[f] for f in fathers]
+                features = [
+                    self.estimator.indexed_features_[f] for f in fathers
+                ]
                 # Update the discretization of the feature
-                res[:, idx] = self.discretizer_.join_fit(
-                    target=idx, features=features, data=self.Xd
+                self.Xd[:, idx] = self.discretizer_.join_fit(
+                    # each feature has to use previous discretization data=res
+                    target=idx,
+                    features=features,
+                    data=self.Xd,
                 )
-                # print(self.discretizer.y_join[:5])
-                upgraded = True
-        return upgraded, res
+                upgrade = True
+        if upgrade:
+            # Update the dataset
+            self.estimator.X_ = self.Xd
+            self.estimator.build_dataset()
+            self.state_names_ = {
+                key: self.discretizer_.get_states_feature(value)
+                for key, value in self.estimator.indexed_features_.items()
+            }
+            states = {"state_names": self.state_names_}
+            # Update the model
+            self.estimator.model_.fit(
+                self.estimator.dataset_,
+                estimator=BayesianEstimator,
+                prior_type="K2",
+                **states,
+            )
 
     # def check_integrity(self, source, X):
     #     # print(f"Checking integrity of {source} data")

bayesclass/test.py

@@ -0,0 +1,19 @@
+from bayesclass.clfs import AODENew, TANNew, KDBNew, AODE
+from benchmark.datasets import Datasets
+import os
+
+os.chdir("../discretizbench")
+dt = Datasets()
+clfan = AODENew()
+clftn = TANNew()
+clfkn = KDBNew()
+# clfa = AODE()
+X, y = dt.load("iris")
+# clfa.fit(X, y)
+clfan.fit(X, y)
+clftn.fit(X, y)
+clfkn.fit(X, y)
+
+
+self.discretizer_.target_
+self.estimator.indexed_features_

bayesclass/tests/conftest.py

@@ -0,0 +1,38 @@
+import pytest
+from sklearn.datasets import load_iris
+from fimdlp.mdlp import FImdlp
+
+
+@pytest.fixture
+def iris():
+    dataset = load_iris()
+    X = dataset["data"]
+    y = dataset["target"]
+    features = dataset["feature_names"]
+    # To make iris dataset has the same values as our iris.arff dataset
+    patch = {(34, 3): (0.2, 0.1), (37, 1): (3.6, 3.1), (37, 2): (1.4, 1.5)}
+    for key, value in patch.items():
+        X[key] = value[1]
+    return X, y, features
+
+
+@pytest.fixture
+def data(iris):
+    return iris[0], iris[1]
+
+
+@pytest.fixture
+def features(iris):
+    return iris[2]
+
+
+@pytest.fixture
+def class_name():
+    return "class"
+
+
+@pytest.fixture
+def data_disc(data):
+    clf = FImdlp()
+    X, y = data
+    return clf.fit_transform(X, y), y

bayesclass/tests/test_AODE.py

@@ -1,6 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
 from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
@@ -10,26 +9,19 @@ from bayesclass.clfs import AODE
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return AODE(random_state=17)
 
 
-def test_AODE_default_hyperparameters(data, clf):
+def test_AODE_default_hyperparameters(data_disc, clf):
     # Test default values of hyperparameters
     assert not clf.show_progress
     assert clf.random_state == 17
     clf = AODE(show_progress=True)
     assert clf.show_progress
     assert clf.random_state is None
-    clf.fit(*data)
+    clf.fit(*data_disc)
     assert clf.class_name_ == "class"
     assert clf.feature_names_in_ == [
         "feature_0",
@@ -42,37 +34,35 @@ def test_AODE_default_hyperparameters(data, clf):
 @image_comparison(
     baseline_images=["line_dashes_AODE"], remove_text=True, extensions=["png"]
 )
-def test_AODE_plot(data, clf):
+def test_AODE_plot(data_disc, features, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data_disc, features=features)
     clf.plot("AODE Iris")
 
 
-def test_AODE_version(clf, data):
+def test_AODE_version(clf, features, data_disc):
     """Check AODE version."""
     assert __version__ == clf.version()
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data_disc, features=features)
     assert __version__ == clf.version()
 
 
-def test_AODE_nodes_edges(clf, data):
+def test_AODE_nodes_edges(clf, data_disc):
     assert clf.nodes_edges() == (0, 0)
-    clf.fit(*data)
+    clf.fit(*data_disc)
     assert clf.nodes_leaves() == (20, 28)
 
 
-def test_AODE_states(clf, data):
+def test_AODE_states(clf, data_disc):
     assert clf.states_ == 0
-    clf.fit(*data)
-    assert clf.states_ == 23
+    clf.fit(*data_disc)
+    assert clf.states_ == 19
     assert clf.depth_ == clf.states_
 
 
-def test_AODE_classifier(data, clf):
-    clf.fit(*data)
+def test_AODE_classifier(data_disc, clf):
+    clf.fit(*data_disc)
     attribs = [
         "feature_names_in_",
         "class_name_",
@@ -82,28 +72,28 @@ def test_AODE_classifier(data, clf):
     ]
     for attr in attribs:
         assert hasattr(clf, attr)
-    X = data[0]
-    y = data[1]
+    X = data_disc[0]
+    y = data_disc[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 147
+    assert sum(y == y_pred) == 146
 
 
-def test_AODE_wrong_num_features(data, clf):
+def test_AODE_wrong_num_features(data_disc, clf):
     with pytest.raises(
         ValueError,
         match="Number of features does not match the number of columns in X",
     ):
-        clf.fit(*data, features=["feature_1", "feature_2"])
+        clf.fit(*data_disc, features=["feature_1", "feature_2"])
 
 
-def test_AODE_wrong_hyperparam(data, clf):
+def test_AODE_wrong_hyperparam(data_disc, clf):
     with pytest.raises(ValueError, match="Unexpected argument: wrong_param"):
-        clf.fit(*data, wrong_param="wrong_param")
+        clf.fit(*data_disc, wrong_param="wrong_param")
 
 
-def test_AODE_error_size_predict(data, clf):
-    X, y = data
+def test_AODE_error_size_predict(data_disc, clf):
+    X, y = data_disc
     clf.fit(X, y)
     with pytest.raises(ValueError):
         X_diff_size = np.ones((10, X.shape[1] + 1))

bayesclass/tests/test_AODENew.py

@@ -1,7 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
-from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
 
@@ -10,13 +8,6 @@ from bayesclass.clfs import AODENew
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return AODENew(random_state=17)
@@ -44,19 +35,17 @@ def test_AODENew_default_hyperparameters(data, clf):
     remove_text=True,
     extensions=["png"],
 )
-def test_AODENew_plot(data, clf):
+def test_AODENew_plot(data, features, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data, features=features)
     clf.plot("AODE Iris")
 
 
 def test_AODENew_version(clf, data):
     """Check AODENew version."""
     assert __version__ == clf.version()
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data)
     assert __version__ == clf.version()
 
 
@@ -69,7 +58,7 @@ def test_AODENew_nodes_edges(clf, data):
 def test_AODENew_states(clf, data):
     assert clf.states_ == 0
     clf.fit(*data)
-    assert clf.states_ == 22.75
+    assert clf.states_ == 17.75
     assert clf.depth_ == clf.states_
 
 
@@ -88,17 +77,17 @@ def test_AODENew_classifier(data, clf):
     y = data[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 147
+    assert sum(y == y_pred) == 146
 
 
-def test_AODENew_local_discretization(clf, data):
+def test_AODENew_local_discretization(clf, data_disc):
     expected_data = [
         [-1, [0, -1], [0, -1], [0, -1]],
         [[1, -1], -1, [1, -1], [1, -1]],
         [[2, -1], [2, -1], -1, [2, -1]],
         [[3, -1], [3, -1], [3, -1], -1],
     ]
-    clf.fit(*data)
+    clf.fit(*data_disc)
     for idx, estimator in enumerate(clf.estimators_):
         expected = expected_data[idx]
         for feature in range(4):

bayesclass/tests/test_KDB.py

@@ -1,6 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
 from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
@@ -11,19 +10,12 @@ from bayesclass.clfs import KDB
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return KDB(k=3)
 
 
-def test_KDB_default_hyperparameters(data, clf):
+def test_KDB_default_hyperparameters(data_disc, clf):
     # Test default values of hyperparameters
     assert not clf.show_progress
     assert clf.random_state is None
@@ -32,7 +24,7 @@ def test_KDB_default_hyperparameters(data, clf):
     assert clf.show_progress
     assert clf.random_state == 17
     assert clf.k == 3
-    clf.fit(*data)
+    clf.fit(*data_disc)
     assert clf.class_name_ == "class"
     assert clf.feature_names_in_ == [
         "feature_0",
@@ -47,57 +39,56 @@ def test_KDB_version(clf):
     assert __version__ == clf.version()
 
 
-def test_KDB_nodes_edges(clf, data):
+def test_KDB_nodes_edges(clf, data_disc):
     assert clf.nodes_edges() == (0, 0)
-    clf.fit(*data)
-    assert clf.nodes_leaves() == (5, 10)
+    clf.fit(*data_disc)
+    assert clf.nodes_leaves() == (5, 9)
 
 
-def test_KDB_states(clf, data):
+def test_KDB_states(clf, data_disc):
     assert clf.states_ == 0
-    clf.fit(*data)
-    assert clf.states_ == 23
+    clf.fit(*data_disc)
+    assert clf.states_ == 19
     assert clf.depth_ == clf.states_
 
 
-def test_KDB_classifier(data, clf):
-    clf.fit(*data)
+def test_KDB_classifier(data_disc, clf):
+    clf.fit(*data_disc)
     attribs = ["classes_", "X_", "y_", "feature_names_in_", "class_name_"]
     for attr in attribs:
         assert hasattr(clf, attr)
-    X = data[0]
-    y = data[1]
+    X = data_disc[0]
+    y = data_disc[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 148
+    assert sum(y == y_pred) == 146
 
 
 @image_comparison(
     baseline_images=["line_dashes_KDB"], remove_text=True, extensions=["png"]
 )
-def test_KDB_plot(data, clf):
+def test_KDB_plot(data_disc, features, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data_disc, features=features)
     clf.plot("KDB Iris")
 
 
-def test_KDB_wrong_num_features(data, clf):
+def test_KDB_wrong_num_features(data_disc, clf):
     with pytest.raises(
         ValueError,
         match="Number of features does not match the number of columns in X",
     ):
-        clf.fit(*data, features=["feature_1", "feature_2"])
+        clf.fit(*data_disc, features=["feature_1", "feature_2"])
 
 
-def test_KDB_wrong_hyperparam(data, clf):
+def test_KDB_wrong_hyperparam(data_disc, clf):
     with pytest.raises(ValueError, match="Unexpected argument: wrong_param"):
-        clf.fit(*data, wrong_param="wrong_param")
+        clf.fit(*data_disc, wrong_param="wrong_param")
 
 
-def test_KDB_error_size_predict(data, clf):
-    X, y = data
+def test_KDB_error_size_predict(data_disc, clf):
+    X, y = data_disc
     clf.fit(X, y)
     with pytest.raises(ValueError):
         X_diff_size = np.ones((10, X.shape[1] + 1))

bayesclass/tests/test_KDBNew.py

@@ -1,7 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
-from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
 from pgmpy.models import BayesianNetwork
@@ -11,13 +9,6 @@ from bayesclass.clfs import KDBNew
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return KDBNew(k=3)
@@ -50,13 +41,13 @@ def test_KDBNew_version(clf):
 def test_KDBNew_nodes_edges(clf, data):
     assert clf.nodes_edges() == (0, 0)
     clf.fit(*data)
-    assert clf.nodes_leaves() == (5, 10)
+    assert clf.nodes_leaves() == (5, 9)
 
 
 def test_KDBNew_states(clf, data):
     assert clf.states_ == 0
     clf.fit(*data)
-    assert clf.states_ == 23
+    assert clf.states_ == 22
     assert clf.depth_ == clf.states_
 
 
@@ -69,14 +60,15 @@ def test_KDBNew_classifier(data, clf):
     y = data[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 148
+    assert sum(y == y_pred) == 145
 
 
 def test_KDBNew_local_discretization(clf, data):
-    expected = [[1, -1], -1, [0, 1, 3, -1], [1, 0, -1]]
+    expected = [[1, -1], -1, [0, 1, 3, -1], [1, -1]]
     clf.fit(*data)
     for feature in range(4):
         computed = clf.estimator_.discretizer_.target_[feature]
+        print("computed:", computed)
         if type(computed) == list:
             for j, k in zip(expected[feature], computed):
                 assert j == k
@@ -92,11 +84,10 @@ def test_KDBNew_local_discretization(clf, data):
     remove_text=True,
     extensions=["png"],
 )
-def test_KDBNew_plot(data, clf):
+def test_KDBNew_plot(data, features, class_name, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"])
+    clf.fit(*data, features=features, class_name=class_name)
     clf.plot("KDBNew Iris")
 
 

bayesclass/tests/test_TAN.py

@@ -1,7 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
-from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
 
@@ -10,26 +8,19 @@ from bayesclass.clfs import TAN
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return TAN(random_state=17)
 
 
-def test_TAN_default_hyperparameters(data, clf):
+def test_TAN_default_hyperparameters(data_disc, clf):
     # Test default values of hyperparameters
     assert not clf.show_progress
     assert clf.random_state == 17
     clf = TAN(show_progress=True)
     assert clf.show_progress
     assert clf.random_state is None
-    clf.fit(*data)
+    clf.fit(*data_disc)
     assert clf.head_ == 0
     assert clf.class_name_ == "class"
     assert clf.feature_names_in_ == [
@@ -45,26 +36,26 @@ def test_TAN_version(clf):
     assert __version__ == clf.version()
 
 
-def test_TAN_nodes_edges(clf, data):
+def test_TAN_nodes_edges(clf, data_disc):
     assert clf.nodes_edges() == (0, 0)
-    clf.fit(*data, head="random")
+    clf.fit(*data_disc, head="random")
     assert clf.nodes_leaves() == (5, 7)
 
 
-def test_TAN_states(clf, data):
+def test_TAN_states(clf, data_disc):
     assert clf.states_ == 0
-    clf.fit(*data)
-    assert clf.states_ == 23
+    clf.fit(*data_disc)
+    assert clf.states_ == 19
     assert clf.depth_ == clf.states_
 
 
-def test_TAN_random_head(clf, data):
-    clf.fit(*data, head="random")
+def test_TAN_random_head(clf, data_disc):
+    clf.fit(*data_disc, head="random")
     assert clf.head_ == 3
 
 
-def test_TAN_classifier(data, clf):
-    clf.fit(*data)
+def test_TAN_classifier(data_disc, clf):
+    clf.fit(*data_disc)
     attribs = [
         "classes_",
         "X_",
@@ -75,44 +66,43 @@ def test_TAN_classifier(data, clf):
     ]
     for attr in attribs:
         assert hasattr(clf, attr)
-    X = data[0]
-    y = data[1]
+    X = data_disc[0]
+    y = data_disc[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 147
+    assert sum(y == y_pred) == 146
 
 
 @image_comparison(
     baseline_images=["line_dashes_TAN"], remove_text=True, extensions=["png"]
 )
-def test_TAN_plot(data, clf):
+def test_TAN_plot(data_disc, features, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"], head=0)
+    clf.fit(*data_disc, features=features, head=0)
     clf.plot("TAN Iris head=0")
 
 
-def test_TAN_wrong_num_features(data, clf):
+def test_TAN_wrong_num_features(data_disc, clf):
     with pytest.raises(
         ValueError,
         match="Number of features does not match the number of columns in X",
     ):
-        clf.fit(*data, features=["feature_1", "feature_2"])
+        clf.fit(*data_disc, features=["feature_1", "feature_2"])
 
 
-def test_TAN_wrong_hyperparam(data, clf):
+def test_TAN_wrong_hyperparam(data_disc, clf):
     with pytest.raises(ValueError, match="Unexpected argument: wrong_param"):
-        clf.fit(*data, wrong_param="wrong_param")
+        clf.fit(*data_disc, wrong_param="wrong_param")
 
 
-def test_TAN_head_out_of_range(data, clf):
+def test_TAN_head_out_of_range(data_disc, clf):
     with pytest.raises(ValueError, match="Head index out of range"):
-        clf.fit(*data, head=4)
+        clf.fit(*data_disc, head=4)
 
 
-def test_TAN_error_size_predict(data, clf):
-    X, y = data
+def test_TAN_error_size_predict(data_disc, clf):
+    X, y = data_disc
     clf.fit(X, y)
     with pytest.raises(ValueError):
         X_diff_size = np.ones((10, X.shape[1] + 1))

bayesclass/tests/test_TANNew.py

@@ -1,7 +1,5 @@
 import pytest
 import numpy as np
-from sklearn.datasets import load_iris
-from sklearn.preprocessing import KBinsDiscretizer
 from matplotlib.testing.decorators import image_comparison
 from matplotlib.testing.conftest import mpl_test_settings
 
@@ -10,13 +8,6 @@ from bayesclass.clfs import TANNew
 from .._version import __version__
 
 
-@pytest.fixture
-def data():
-    X, y = load_iris(return_X_y=True)
-    enc = KBinsDiscretizer(encode="ordinal")
-    return enc.fit_transform(X), y
-
-
 @pytest.fixture
 def clf():
     return TANNew(random_state=17)
@@ -54,7 +45,7 @@ def test_TANNew_nodes_edges(clf, data):
 def test_TANNew_states(clf, data):
     assert clf.states_ == 0
     clf.fit(*data)
-    assert clf.states_ == 22
+    assert clf.states_ == 18
     assert clf.depth_ == clf.states_
 
 
@@ -88,7 +79,7 @@ def test_TANNew_classifier(data, clf):
     y = data[1]
     y_pred = clf.predict(X)
     assert y_pred.shape == (X.shape[0],)
-    assert sum(y == y_pred) == 145
+    assert sum(y == y_pred) == 146
 
 
 @image_comparison(
@@ -96,11 +87,10 @@ def test_TANNew_classifier(data, clf):
     remove_text=True,
     extensions=["png"],
 )
-def test_TANNew_plot(data, clf):
+def test_TANNew_plot(data, features, clf):
     # mpl_test_settings will automatically clean these internal side effects
     mpl_test_settings
-    dataset = load_iris(as_frame=True)
-    clf.fit(*data, features=dataset["feature_names"], head=0)
+    clf.fit(*data, features=features, head=0)
     clf.plot("TANNew Iris head=0")