FIX: crf compatibility with method and crf intervals shape

mapie/conformity_scores/residual_conformity_scores.py

@@ -433,7 +433,7 @@ def get_estimation_distribution(
  ``conformity_scores`` can be either the conformity scores or
  the quantile of the conformity scores.
  """
- r_pred = self._predict_residual_estimator(X)
+ r_pred = self._predict_residual_estimator(X).reshape((-1, 1))
  if not self.prefit:
  return np.add(
  y_pred,

mapie/regression/regression.py

@@ -71,8 +71,9 @@ class MapieRegressor(BaseEstimator, RegressorMixin):
  - ``"split"``, does not involve cross-validation but a division
  of the data into training and calibration subsets. The splitter
  used is the following: ``sklearn.model_selection.ShuffleSplit``.
+ ``method`` parameter is set to ``"base"``.
  - ``"prefit"``, assumes that ``estimator`` has been fitted already,
- and the ``method`` parameter is ignored.
+ and the ``method`` parameter is set to ``"base"``.
  All data provided in the ``fit`` method is then used
  for computing conformity scores only.
  At prediction time, quantiles of these conformity scores are used
@@ -407,13 +408,19 @@ def _check_fit_parameters(
  ------
  ValueError
  If conformity score is FittedResidualNormalizing score and method
- is neither ``"prefit"`` or ``"split"``
+ is neither ``"prefit"`` or ``"split"``.
+
+ ValueError
+ If ``cv`` is `"prefit"`` or ``"split"`` and ``method`` is not
+ ``"base"``.
  """
  # Checking
  self._check_parameters()
  cv = check_cv(
  self.cv, test_size=self.test_size, random_state=self.random_state
  )
+ if self.cv in ["split", "prefit"] and self.method != "base":
+ self.method = "base"
  estimator = self._check_estimator(self.estimator)
  agg_function = self._check_agg_function(self.agg_function)
  cs_estimator = check_conformity_score(
@@ -422,8 +429,8 @@ def _check_fit_parameters(
  if isinstance(cs_estimator, ConformalResidualFittingScore) and \
  self.cv not in ["split", "prefit"]:
  raise ValueError(
- "The FittedResisualNormalizingScore can be used only with "
- "``split`` and ``prefit`` methods"
+ "The ConformalResidualFittingScore can be used only with "
+ "``cv='split'`` and ``cv='prefit'``"
  )
 
  X, y = indexable(X, y)

mapie/tests/test_conformity_scores.py

@@ -1,3 +1,5 @@
+from typing import Any
+
 import numpy as np
 import pytest
 
@@ -372,3 +374,23 @@ def test_crf_prefit_get_estimation_distribution() -> None:
  crf_conf_score.get_estimation_distribution(
  X_toy, y_pred_list, conf_scores
  )
+
+
+@pytest.mark.parametrize("score", [AbsoluteConformityScore(),
+ GammaConformityScore(),
+ ConformalResidualFittingScore()])
+@pytest.mark.parametrize("alpha", [[0.3], [0.5, 0.4]])
+def test_intervals_shape_with_every_score(
+ score: ConformityScore,
+ alpha: Any
+) -> None:
+ mapie_reg = MapieRegressor(
+ method="base", cv="split", conformity_score=score
+ )
+ X = np.concatenate((X_toy, X_toy))
+ y = np.concatenate((y_toy, y_toy))
+ mapie_reg = mapie_reg.fit(X, y)
+ y_pred, intervals = mapie_reg.predict(X, alpha=alpha)
+ n_samples = X.shape[0]
+ assert y_pred.shape[0] == n_samples
+ assert intervals.shape == (n_samples, 2, len(alpha))