UPD: dosctrings + beautify

mapie/conformity_scores/conformity_scores.py

@@ -69,18 +69,18 @@ def get_signed_conformity_scores(
 
  Parameters
  ----------
- X: ArrayLike of shape (n_samples_calib, n_features)
+ X: ArrayLike of shape (n_samples, n_features)
  Observed feature values.
 
- y: ArrayLike of shape (n_samples_calib,)
+ y: ArrayLike of shape (n_samples,)
  Observed target values.
 
- y_pred: ArrayLike of shape (n_samples_calib,)
+ y_pred: ArrayLike of shape (n_samples,)
  Predicted target values.
 
  Returns
  -------
- NDArray of shape (n_samples_calib,)
+ NDArray of shape (n_samples,)
  Signed conformity scores.
  """
 
@@ -101,24 +101,24 @@ def get_estimation_distribution(
 
  Parameters
  ----------
- X: ArrayLike of shape (n_samples_calib, n_features)
+ X: ArrayLike of shape (n_samples, n_features)
  Observed feature values.
 
  y_pred: ArrayLike
- The shape is either (n_samples_calib, n_samples_train): when the
+ The shape is either (n_samples, n_references): when the
  method is called in ``get_bounds`` it needs a prediction per train
  sample for each test sample to compute the bounds.
- Or (n_samples_calib, 1): when it is called in ``check_consistency``
+ Or (n_samples,): when it is called in ``check_consistency``
 
  conformity_scores: ArrayLike
- The shape is either (n_samples_calib, n_alpha) when it is the
- conformity scores themselves or (n_alpha, 1) when it is only the
+ The shape is either (n_samples, 1) when it is the
+ conformity scores themselves or (1, n_alpha) when it is only the
  quantile of the conformity scores.
 
  Returns
  -------
- NDArray of shape (n_samples_calib, n_alpha) or
- (n_samples_calib, n_samples_train) according to the shape of ``y_pred``
+ NDArray of shape (n_samples, n_alpha) or
+ (n_samples, n_references) according to the shape of ``y_pred``
  Observed values.
  """
 
@@ -142,16 +142,16 @@ def check_consistency(
 
  Parameters
  ----------
- X: ArrayLike of shape (n_samples_calib, n_features)
+ X: ArrayLike of shape (n_samples, n_features)
  Observed feature values.
 
- y: ArrayLike of shape (n_samples_calib,)
+ y: ArrayLike of shape (n_samples,)
  Observed target values.
 
- y_pred: ArrayLike of shape (n_samples_calib,)
+ y_pred: ArrayLike of shape (n_samples,)
  Predicted target values.
 
- conformity_scores: ArrayLike of shape (n_samples_calib,)
+ conformity_scores: ArrayLike of shape (n_samples,)
  Conformity scores.
 
  Raises
@@ -188,18 +188,18 @@ def get_conformity_scores(
 
  Parameters
  ----------
- X: NDArray of shape (n_samples_calib, n_features)
+ X: NDArray of shape (n_samples, n_features)
  Observed feature values.
 
- y: NDArray of shape (n_samples_calib,)
+ y: NDArray of shape (n_samples,)
  Observed target values.
 
- y_pred: NDArray of shape (n_samples_calib,)
+ y_pred: NDArray of shape (n_samples,)
  Predicted target values.
 
  Returns
  -------
- NDArray of shape (n_samples_calib, 1)
+ NDArray of shape (n_samples,)
  Conformity scores.
  """
  conformity_scores = self.get_signed_conformity_scores(X, y, y_pred)
@@ -222,8 +222,8 @@ def get_quantile(
 
  Parameters
  ----------
- values: NDArray of shape (n_samples_calib, n_alpha) or
- (n_samples_calib, n_samples_train)
+ values: NDArray of shape (n_samples,) or
+ (n_samples, n_references)
  Values from which the quantile is computed, it can be the
  conformity scores or the conformity scores aggregated with
  the predictions.
@@ -240,7 +240,7 @@ def get_quantile(
 
  Returns
  -------
- NDArray of shape (n_alpha,)
+ NDArray of shape (1, n_alpha) or (n_samples, n_alpha)
  The quantile of the conformity scores.
  """
  quantile = np.column_stack([
@@ -269,13 +269,13 @@ def get_bounds(
 
  Parameters
  ----------
- X: ArrayLike of shape (n_samples_test, n_features)
+ X: ArrayLike of shape (n_samples, n_features)
  Observed feature values.
 
  estimator: EnsembleEstimator
  Estimator that is fitted to predict y from X.
 
- conformity_scores: ArrayLike of shape (n_samples_calib,)
+ conformity_scores: ArrayLike of shape (n_samples,)
  Conformity scores.
 
  alpha_np: NDArray of shape (n_alpha,)
@@ -295,11 +295,11 @@ def get_bounds(
  Returns
  -------
  Tuple[NDArray, NDArray, NDArray]
- - The predictions itself. (y_pred) of shape (n_samples_test,).
+ - The predictions itself. (y_pred) of shape (n_samples,).
  - The lower bounds of the prediction intervals of shape
- (n_samples_test,).
+ (n_samples, n_alpha).
  - The upper bounds of the prediction intervals of shape
- (n_samples_test,).
+ (n_samples, n_alpha).
  """
  y_pred, y_pred_low, y_pred_up = estimator.predict(X, ensemble)
  signed = -1 if self.sym else 1

mapie/conformity_scores/residual_conformity_scores.py

@@ -29,8 +29,8 @@ def get_signed_conformity_scores(
  y_pred: ArrayLike,
  ) -> NDArray:
  """
- Compute the signed conformity scores from the observed values
- and the predicted ones, from the following formula:
+ Compute the signed conformity scores from the predicted values
+ and the observed ones, from the following formula:
  signed conformity score = y - y_pred
  """
  return np.subtract(y, y_pred)
@@ -43,7 +43,7 @@ def get_estimation_distribution(
  ) -> NDArray:
  """
  Compute samples of the estimation distribution from the predicted
- targets and ``conformity_scores``, from the following formula:
+ values and the conformity scores, from the following formula:
  signed conformity score = y - y_pred
  <=> y = y_pred + signed conformity score
 
@@ -123,7 +123,7 @@ def get_estimation_distribution(
  ) -> NDArray:
  """
  Compute samples of the estimation distribution from the predicted
- targets and ``conformity_scores``, from the following formula:
+ values and the conformity scores, from the following formula:
  signed conformity score = (y - y_pred) / y_pred
  <=> y = y_pred * (1 + signed conformity score)
 

mapie/tests/test_conformity_scores.py

@@ -22,8 +22,9 @@ def get_signed_conformity_scores(
  ) -> NDArray:
  return np.subtract(y, y_pred)
 
- def get_estimation_distribution(self, X: ArrayLike, y_pred: ArrayLike,
- conformity_scores: ArrayLike) -> NDArray:
+ def get_estimation_distribution(
+ self, X: ArrayLike, y_pred: ArrayLike, conformity_scores: ArrayLike
+ ) -> NDArray:
  """
  A positive constant is added to the sum between predictions and
  conformity scores to make the estimated distribution inconsistent
@@ -65,8 +66,9 @@ def test_absolute_conformity_score_get_estimation_distribution(
 ) -> None:
  """Test conformity observed value computation for AbsoluteConformityScore.""" # noqa: E501
  abs_conf_score = AbsoluteConformityScore()
- y_obs = abs_conf_score.get_estimation_distribution(X_toy, y_pred,
- conf_scores)
+ y_obs = abs_conf_score.get_estimation_distribution(
+ X_toy, y_pred, conf_scores
+ )
  np.testing.assert_allclose(y_obs, y_toy)
 
 
@@ -77,8 +79,9 @@ def test_absolute_conformity_score_consistency(y_pred: NDArray) -> None:
  signed_conf_scores = abs_conf_score.get_signed_conformity_scores(
  X_toy, y_toy, y_pred
  )
- y_obs = abs_conf_score.get_estimation_distribution(X_toy, y_pred,
- signed_conf_scores)
+ y_obs = abs_conf_score.get_estimation_distribution(
+ X_toy, y_pred, signed_conf_scores
+ )
  np.testing.assert_allclose(y_obs, y_toy)
 
 
@@ -89,7 +92,8 @@ def test_gamma_conformity_score_get_conformity_scores(
  """Test conformity score computation for GammaConformityScore."""
  gamma_conf_score = GammaConformityScore()
  conf_scores = gamma_conf_score.get_conformity_scores(
- X_toy, y_toy, y_pred)
+ X_toy, y_toy, y_pred
+ )
  expected_signed_conf_scores = np.array(conf_scores_gamma_list)
  np.testing.assert_allclose(conf_scores, expected_signed_conf_scores)
 
@@ -107,8 +111,9 @@ def test_gamma_conformity_score_get_estimation_distribution(
 ) -> None:
  """Test conformity observed value computation for GammaConformityScore.""" # noqa: E501
  gamma_conf_score = GammaConformityScore()
- y_obs = gamma_conf_score.get_estimation_distribution(X_toy, y_pred,
- conf_scores)
+ y_obs = gamma_conf_score.get_estimation_distribution(
+ X_toy, y_pred, conf_scores
+ )
  np.testing.assert_allclose(y_obs, y_toy)
 
 
@@ -119,8 +124,9 @@ def test_gamma_conformity_score_consistency(y_pred: NDArray) -> None:
  signed_conf_scores = gamma_conf_score.get_signed_conformity_scores(
  X_toy, y_toy, y_pred
  )
- y_obs = gamma_conf_score.get_estimation_distribution(X_toy, y_pred,
- signed_conf_scores)
+ y_obs = gamma_conf_score.get_estimation_distribution(
+ X_toy, y_pred, signed_conf_scores
+ )
  np.testing.assert_allclose(y_obs, y_toy)
 
 
@@ -183,14 +189,15 @@ def test_gamma_conformity_score_check_predicted_value(
  ValueError,
  match=r".*At least one of the predicted target is negative.*"
  ):
- gamma_conf_score.get_estimation_distribution(X_toy, y_pred,
- conf_scores)
+ gamma_conf_score.get_estimation_distribution(
+ X_toy, y_pred, conf_scores
+ )
 
 
 def test_check_consistency() -> None:
  """
- Test that a dummy ConformityScore class that gives inconsistent
- conformity scores and distributions raises an error.
+ Test that a dummy ConformityScore class that gives inconsistent scores
+ and distributions raises an error.
  """
  dummy_conf_score = DummyConformityScore()
  conformity_scores = dummy_conf_score.get_signed_conformity_scores(