test: Adapt coefficients() logic for meta-estimators (#1616)

followup for PR #1575

Author: Schefflera-Arboricola

skore/src/skore/sklearn/_estimator/feature_importance_accessor.py

@@ -206,8 +206,12 @@ def coefficients(self) -> pd.DataFrame:
         try:
             intercept = np.atleast_2d(estimator.intercept_)
         except AttributeError:
+            # TransformedTargetRegressor() does not expose `intercept_`
+            intercept = np.atleast_2d(estimator.regressor_.intercept_)
+            # Uncomment when SGDOneClassSVM is fully supported by EstimatorReport
+            # except AttributeError:
             # SGDOneClassSVM does not expose `intercept_`
-            intercept = None
+            # intercept = None
 
         try:
             coef = np.atleast_2d(estimator.coef_)

skore/tests/unit/sklearn/estimator/feature_importance/test_coefficients.py

@@ -224,19 +224,15 @@ def test_all_sklearn_estimators(
     report = EstimatorReport(estimator)
     result = report.feature_importance.coefficients()
 
-    rows = [
+    assert result.shape == (6, 1)
+    assert result.index.tolist() == [
         "Intercept",
         "Feature #0",
         "Feature #1",
         "Feature #2",
         "Feature #3",
         "Feature #4",
     ]
-    if result.shape == (5, 1):  # for TransformedTargetRegressor()
-        assert rows[1:] == result.index.tolist()
-    else:
-        assert result.shape == (6, 1)
-        assert rows == result.index.tolist()
 
     assert result.columns.tolist() == ["Coefficient"]