feat: return TaggedTable from predict (#73)

### Summary of Changes

The `predict` method returns the original feature vector combined with
the predicted target values. This is exactly what a `TaggedTable` can
express, so we return this instead of a `Table`.

---------

Co-authored-by: lars-reimann <lars-reimann@users.noreply.github.com>

src/safeds/ml/_util_sklearn.py

@@ -6,7 +6,7 @@
 
 
 # noinspection PyProtectedMember
-def fit(model: Any, tagged_table: TaggedTable) -> str:
+def fit(model: Any, tagged_table: TaggedTable) -> None:
     """
     Fit a model for a given tagged table.
 
@@ -17,11 +17,6 @@ def fit(model: Any, tagged_table: TaggedTable) -> str:
     tagged_table : TaggedTable
         The tagged table containing the feature and target vectors.
 
-    Returns
-    -------
-    target_name : str
-        The target column name, inferred from the tagged table.
-
     Raises
     ------
     LearningError
@@ -32,15 +27,14 @@ def fit(model: Any, tagged_table: TaggedTable) -> str:
             tagged_table.feature_vectors._data,
             tagged_table.target_values._data,
         )
-        return tagged_table.target_values.name
     except ValueError as exception:
         raise LearningError(str(exception)) from exception
     except Exception as exception:
         raise LearningError(None) from exception
 
 
 # noinspection PyProtectedMember
-def predict(model: Any, dataset: Table, target_name: str) -> Table:
+def predict(model: Any, dataset: Table, target_name: str) -> TaggedTable:
     """
     Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -73,7 +67,7 @@ def predict(model: Any, dataset: Table, target_name: str) -> Table:
                 f"Dataset already contains '{target_name}' column. Please rename this column"
             )
         result_set[target_name] = predicted_target_vector
-        return Table(result_set)
+        return TaggedTable(Table(result_set), target_column=target_name)
     except NotFittedError as exception:
         raise PredictionError("The model was not trained") from exception
     except ValueError as exception:

src/safeds/ml/classification/_ada_boost.py

@@ -1,6 +1,6 @@
 # noinspection PyProtectedMember
-import safeds.ml._util_sklearn
 from safeds.data.tabular.containers import Table, TaggedTable
+from safeds.ml._util_sklearn import fit, predict
 from sklearn.ensemble import AdaBoostClassifier as sk_AdaBoostClassifier
 
 from ._classifier import Classifier
@@ -14,8 +14,8 @@ class AdaBoost(Classifier):
     """
 
     def __init__(self) -> None:
-        self._classification = sk_AdaBoostClassifier()
-        self.target_name = ""
+        self._wrapped_classifier = sk_AdaBoostClassifier()
+        self._target_name = ""
 
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -31,11 +31,10 @@ def fit(self, training_set: TaggedTable) -> None:
         LearningError
             If the tagged table contains invalid values or if the training failed.
         """
-        self.target_name = safeds.ml._util_sklearn.fit(
-            self._classification, training_set
-        )
+        fit(self._wrapped_classifier, training_set)
+        self._target_name = training_set.target_values.name
 
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -46,16 +45,12 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises
         ------
         PredictionError
             If prediction with the given dataset failed.
         """
-        return safeds.ml._util_sklearn.predict(
-            self._classification,
-            dataset,
-            self.target_name,
-        )
+        return predict(self._wrapped_classifier, dataset, self._target_name)

src/safeds/ml/classification/_classifier.py

@@ -4,6 +4,10 @@
 
 
 class Classifier(ABC):
+    """
+    Abstract base class for all classifiers.
+    """
+
     @abstractmethod
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -21,7 +25,7 @@ def fit(self, training_set: TaggedTable) -> None:
         """
 
     @abstractmethod
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -32,7 +36,7 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises

src/safeds/ml/classification/_decision_tree.py

@@ -1,6 +1,6 @@
 # noinspection PyProtectedMember
-import safeds.ml._util_sklearn
 from safeds.data.tabular.containers import Table, TaggedTable
+from safeds.ml._util_sklearn import fit, predict
 from sklearn.tree import DecisionTreeClassifier as sk_DecisionTreeClassifier
 
 from ._classifier import Classifier
@@ -14,8 +14,8 @@ class DecisionTree(Classifier):
     """
 
     def __init__(self) -> None:
-        self._classification = sk_DecisionTreeClassifier()
-        self.target_name = ""
+        self._wrapped_classifier = sk_DecisionTreeClassifier()
+        self._target_name = ""
 
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -31,11 +31,10 @@ def fit(self, training_set: TaggedTable) -> None:
         LearningError
             If the tagged table contains invalid values or if the training failed.
         """
-        self.target_name = safeds.ml._util_sklearn.fit(
-            self._classification, training_set
-        )
+        fit(self._wrapped_classifier, training_set)
+        self._target_name = training_set.target_values.name
 
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -46,16 +45,12 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises
         ------
         PredictionError
             If prediction with the given dataset failed.
         """
-        return safeds.ml._util_sklearn.predict(
-            self._classification,
-            dataset,
-            self.target_name,
-        )
+        return predict(self._wrapped_classifier, dataset, self._target_name)

src/safeds/ml/classification/_gradient_boosting_classification.py

@@ -1,7 +1,7 @@
 # noinspection PyProtectedMember
-import safeds.ml._util_sklearn
 from safeds.data.tabular.containers import Table, TaggedTable
-from sklearn.ensemble import GradientBoostingClassifier
+from safeds.ml._util_sklearn import fit, predict
+from sklearn.ensemble import GradientBoostingClassifier as sk_GradientBoostingClassifier
 
 from ._classifier import Classifier
 
@@ -14,8 +14,8 @@ class GradientBoosting(Classifier):
     """
 
     def __init__(self) -> None:
-        self._classification = GradientBoostingClassifier()
-        self.target_name = ""
+        self._wrapped_classifier = sk_GradientBoostingClassifier()
+        self._target_name = ""
 
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -31,12 +31,11 @@ def fit(self, training_set: TaggedTable) -> None:
         LearningError
             If the tagged table contains invalid values or if the training failed.
         """
-        self.target_name = safeds.ml._util_sklearn.fit(
-            self._classification, training_set
-        )
+        fit(self._wrapped_classifier, training_set)
+        self._target_name = training_set.target_values.name
 
     # noinspection PyProtectedMember
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -47,16 +46,12 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises
         ------
         PredictionError
             If prediction with the given dataset failed.
         """
-        return safeds.ml._util_sklearn.predict(
-            self._classification,
-            dataset,
-            self.target_name,
-        )
+        return predict(self._wrapped_classifier, dataset, self._target_name)

src/safeds/ml/classification/_k_nearest_neighbors.py

@@ -1,7 +1,7 @@
 # noinspection PyProtectedMember
-import safeds.ml._util_sklearn
 from safeds.data.tabular.containers import Table, TaggedTable
-from sklearn.neighbors import KNeighborsClassifier
+from safeds.ml._util_sklearn import fit, predict
+from sklearn.neighbors import KNeighborsClassifier as sk_KNeighborsClassifier
 
 from ._classifier import Classifier
 
@@ -18,8 +18,10 @@ class KNearestNeighbors(Classifier):
     """
 
     def __init__(self, n_neighbors: int) -> None:
-        self._classification = KNeighborsClassifier(n_jobs=-1, n_neighbors=n_neighbors)
-        self.target_name = ""
+        self._wrapped_classifier = sk_KNeighborsClassifier(
+            n_jobs=-1, n_neighbors=n_neighbors
+        )
+        self._target_name = ""
 
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -35,11 +37,10 @@ def fit(self, training_set: TaggedTable) -> None:
         LearningError
             If the tagged table contains invalid values or if the training failed.
         """
-        self.target_name = safeds.ml._util_sklearn.fit(
-            self._classification, training_set
-        )
+        fit(self._wrapped_classifier, training_set)
+        self._target_name = training_set.target_values.name
 
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first
 
@@ -50,16 +51,16 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises
         ------
         PredictionError
             If prediction with the given dataset failed.
         """
-        return safeds.ml._util_sklearn.predict(
-            self._classification,
+        return predict(
+            self._wrapped_classifier,
             dataset,
-            self.target_name,
+            self._target_name,
         )

src/safeds/ml/classification/_logistic_regression.py

@@ -1,6 +1,6 @@
 # noinspection PyProtectedMember
-import safeds.ml._util_sklearn
 from safeds.data.tabular.containers import Table, TaggedTable
+from safeds.ml._util_sklearn import fit, predict
 from sklearn.linear_model import LogisticRegression as sk_LogisticRegression
 
 from ._classifier import Classifier
@@ -14,8 +14,8 @@ class LogisticRegression(Classifier):
     """
 
     def __init__(self) -> None:
-        self._classification = sk_LogisticRegression(n_jobs=-1)
-        self.target_name = ""
+        self._wrapped_classifier = sk_LogisticRegression(n_jobs=-1)
+        self._target_name = ""
 
     def fit(self, training_set: TaggedTable) -> None:
         """
@@ -31,11 +31,10 @@ def fit(self, training_set: TaggedTable) -> None:
         LearningError
             If the tagged table contains invalid values or if the training failed.
         """
-        self.target_name = safeds.ml._util_sklearn.fit(
-            self._classification, training_set
-        )
+        fit(self._wrapped_classifier, training_set)
+        self._target_name = training_set.target_values.name
 
-    def predict(self, dataset: Table) -> Table:
+    def predict(self, dataset: Table) -> TaggedTable:
         """
         Predict a target vector using a dataset containing feature vectors. The model has to be trained first.
 
@@ -46,16 +45,12 @@ def predict(self, dataset: Table) -> Table:
 
         Returns
         -------
-        table : Table
+        table : TaggedTable
             A dataset containing the given feature vectors and the predicted target vector.
 
         Raises
         ------
         PredictionError
             If prediction with the given dataset failed.
         """
-        return safeds.ml._util_sklearn.predict(
-            self._classification,
-            dataset,
-            self.target_name,
-        )
+        return predict(self._wrapped_classifier, dataset, self._target_name)