feat: Add the data accessor to CrossValidationReport

examples/use_cases/plot_employee_salaries.py

@@ -136,6 +136,25 @@
 # %%
 hgbt_model_report.metrics.summarize().frame()
 
+# %%
+# Similarly to what we saw in the previous section, the
+# :class:`skore.CrossValidationReport` also stores some information about the dataset
+# used.
+
+# %%
+data_display = hgbt_model_report.data.analyze()
+data_display
+
+# %%
+# The display obtained allows for a quick overview with the same HTML-based view
+# as the :class:`skrub.TableReport` we have seen earlier. In addition, you can access
+# a :meth:`skore.TableReportDisplay.plot` method to have a particular focus on one
+# potential analysis. For instance, we can get a figure representing the correlation
+# matrix of the dataset.
+
+# %%
+data_display.plot(kind="corr")
+
 # %%
 # We get the results from some statistical metrics aggregated over the cross-validation
 # splits as well as some performance metrics related to the time it took to train and
@@ -153,25 +172,6 @@
 # The favorability of each metric indicates whether the metric is better
 # when higher or lower.
 
-# %%
-# We also have access to some additional information regarding the dataset used for
-# training and testing of the model, similar to what we have seen in the previous
-# section. For example, let's check some information about the training dataset.
-
-# %%
-train_data_display = hgbt_split_1.data.analyze(data_source="train")
-train_data_display
-
-# %%
-# The display obtained allows for a quick overview with the same HTML-based view
-# as the :class:`skrub.TableReport` we have seen earlier. In addition, you can access
-# a :meth:`skore.TableReportDisplay.plot` method to have a particular focus on one
-# potential analysis. For instance, we can get a figure representing the correlation
-# matrix of the training dataset.
-
-# %%
-train_data_display.plot(kind="corr")
-
 # %%
 # Linear model
 # ============

skore/src/skore/_externals/_skrub_compat.py

@@ -10,6 +10,7 @@
 
 
 if skrub_version < parse_version("0.6.0"):
+    tabular_pipeline = skrub.tabular_learner
 
     @dispatch
     def concat(*dataframes, axis=0):
@@ -40,6 +41,9 @@ def _concat_polars(*dataframes, axis=0):
 
     sbd.concat = concat
 
+else:
+    tabular_pipeline = skrub.tabular_pipeline
+
 
 @dispatch
 def to_frame(col):

skore/src/skore/_sklearn/_cross_validation/__init__.py

@@ -1,4 +1,5 @@
 from skore._externals._pandas_accessors import _register_accessor
+from skore._sklearn._cross_validation.data_accessor import _DataAccessor
 from skore._sklearn._cross_validation.feature_importance_accessor import (
     _FeatureImportanceAccessor,
 )
@@ -8,6 +9,7 @@
 )
 
 _register_accessor("metrics", CrossValidationReport)(_MetricsAccessor)
+_register_accessor("data", CrossValidationReport)(_DataAccessor)
 
 _register_accessor("feature_importance", CrossValidationReport)(
     _FeatureImportanceAccessor

skore/src/skore/_sklearn/_cross_validation/data_accessor.py

@@ -0,0 +1,138 @@
+from typing import Literal
+
+import pandas as pd
+from skrub import _dataframe as sbd
+
+from skore._externals._pandas_accessors import DirNamesMixin
+from skore._sklearn._base import _BaseAccessor
+from skore._sklearn._cross_validation.report import CrossValidationReport
+from skore._sklearn._plot import TableReportDisplay
+
+
+class _DataAccessor(_BaseAccessor[CrossValidationReport], DirNamesMixin):
+    def __init__(self, parent: CrossValidationReport) -> None:
+        super().__init__(parent)
+
+    def _retrieve_data_as_frame(
+        self,
+        with_y: bool,
+    ):
+        """Retrieve data as DataFrame.
+
+        Parameters
+        ----------
+        with_y : bool
+            Whether we should check that `y` is not None.
+
+        Returns
+        -------
+        X : DataFrame
+            The input data.
+
+        y : DataFrame or None
+            The target data.
+        """
+        X = self._parent.X
+        y = self._parent.y
+
+        if not sbd.is_dataframe(X):
+            X = pd.DataFrame(X, columns=[f"Feature {i}" for i in range(X.shape[1])])
+
+        if with_y:
+            if y is None:
+                raise ValueError("y is required when `with_y=True`.")
+
+            if isinstance(y, pd.Series):
+                name = y.name if y.name is not None else "Target"
+                y = y.to_frame(name=name)
+            elif not sbd.is_dataframe(y):
+                if y.ndim == 1:
+                    columns = ["Target"]
+                else:
+                    columns = [f"Target {i}" for i in range(y.shape[1])]
+                y = pd.DataFrame(y, columns=columns)
+
+        return X, y
+
+    def analyze(
+        self,
+        with_y: bool = True,
+        subsample: int | None = None,
+        subsample_strategy: Literal["head", "random"] = "head",
+        seed: int | None = None,
+    ) -> TableReportDisplay:
+        """Plot dataset statistics.
+
+        Parameters
+        ----------
+        with_y : bool, default=True
+            Whether to include the target variable in the analysis. If True, the target
+            variable is concatenated horizontally to the features.
+
+        subsample : int, default=None
+            The number of points to subsample the dataframe hold by the display, using
+            the strategy set by ``subsample_strategy``. It must be a strictly positive
+            integer. If ``None``, no subsampling is applied.
+
+        subsample_strategy : {'head', 'random'}, default='head',
+            The strategy used to subsample the dataframe hold by the display. It only
+            has an effect when ``subsample`` is not None.
+
+            - If ``'head'``: subsample by taking the ``subsample`` first points of the
+              dataframe, similar to Pandas: ``df.head(n)``.
+            - If ``'random'``: randomly subsample the dataframe by using a uniform
+              distribution. The random seed is controlled by ``random_state``.
+
+        seed : int, default=None
+            The random seed to use when randomly subsampling. It only has an effect when
+            ``subsample`` is not None and ``subsample_strategy='random'``.
+
+        Returns
+        -------
+        TableReportDisplay
+            A display object containing the dataset statistics and plots.
+
+        Examples
+        --------
+        >>> from sklearn.datasets import load_breast_cancer
+        >>> from sklearn.linear_model import LogisticRegression
+        >>> from skore import CrossValidationReport
+        >>> X, y = load_breast_cancer(return_X_y=True)
+        >>> classifier = LogisticRegression(max_iter=10_000)
+        >>> report = CrossValidationReport(classifier, X=X, y=y, pos_label=1)
+        >>> report.data.analyze().frame()
+        """
+        if subsample_strategy not in (subsample_strategy_options := ("head", "random")):
+            raise ValueError(
+                f"'subsample_strategy' options are {subsample_strategy_options!r}, got "
+                f"{subsample_strategy}."
+            )
+
+        X, y = self._retrieve_data_as_frame(with_y)
+
+        df = sbd.concat(X, y, axis=1) if with_y else X
+
+        if subsample:
+            if subsample_strategy == "head":
+                df = sbd.head(df, subsample)
+            else:  # subsample_strategy == "random":
+                df = sbd.sample(df, subsample, seed=seed)
+
+        return TableReportDisplay._compute_data_for_display(df)
+
+    ####################################################################################
+    # Methods related to the help tree
+    ####################################################################################
+
+    def _format_method_name(self, name: str) -> str:
+        return f"{name}(...)".ljust(29)
+
+    def _get_help_panel_title(self) -> str:
+        return "[bold cyan]Available data methods[/bold cyan]"
+
+    def _get_help_tree_title(self) -> str:
+        return "[bold cyan]report.data[/bold cyan]"
+
+    def __repr__(self) -> str:
+        """Return a string representation using rich."""
+        return self._rich_repr(class_name="skore.CrossValidationReport.data")

skore/src/skore/project/project.py

@@ -6,7 +6,7 @@
 from importlib.metadata import entry_points
 from typing import Any, Literal
 
-from skore._sklearn._estimator.report import EstimatorReport
+from skore import EstimatorReport
 from skore.project.summary import Summary
 
 
@@ -95,7 +95,7 @@ class Project:
     >>> from sklearn.datasets import make_classification, make_regression
     >>> from sklearn.linear_model import LinearRegression, LogisticRegression
     >>> from sklearn.model_selection import train_test_split
-    >>> from skore._sklearn import EstimatorReport
+    >>> from skore import EstimatorReport
     >>>
     >>> X, y = make_classification(random_state=42)
     >>> X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)