ENH: add random state to test and lint change

mapie/control_risk/ltt.py

@@ -18,7 +18,7 @@ def _ltt_procedure(
  Apply the Learn-Then-Test procedure for risk control.
  This procedure is called in ``MapieMultiLabelClassifier``
  if ``metric=precision``.
- This will apply learn then test procedure for 
+ This will apply learn then test procedure for
  precision control.
  Note that we will do a multiple test for ``r_hat`` that are
  less than level ``alpha_np``.

mapie/multi_label_classification.py

@@ -3,8 +3,6 @@
 import warnings
 from typing import Iterable, Optional, Sequence, Tuple, Union, cast
 
-import itertools
-
 import numpy as np
 from sklearn.base import BaseEstimator, ClassifierMixin
 from sklearn.linear_model import LogisticRegression
@@ -99,7 +97,7 @@ class MapieMultiLabelClassifier(BaseEstimator, ClassifierMixin):
 
  lambdas_star: ArrayLike of shape (n_lambdas)
  Optimal threshold for a given alpha.
- 
+
  valid_index: List[List[Any]]
  List of list of all index that satisfy fwer controlling
  for learn then test procedure. This attribute is compute

mapie/tests/test_control_risk.py

@@ -66,6 +66,12 @@
 
 wrong_delta = None
 
+random_state = 42
+prng = np.random.RandomState(random_state)
+y_1 = prng.random(51)
+y_2 = prng.random((51, 5))
+y_3 = prng.randint(0, 2, 51)
+
 
 def test_compute_recall_equal() -> None:
  """Test that compute_recall give good result"""
@@ -122,7 +128,9 @@ def test_compute_precision_with_wrong_shape() -> None:
  with pytest.raises(ValueError, match=r".*y_pred_proba should be a 3d*"):
  _compute_risk_precision(lambdas, y_preds_proba.squeeze(), y_toy)
  with pytest.raises(ValueError, match=r".*y should be a 2d*"):
- _compute_risk_precision(lambdas, y_preds_proba, np.expand_dims(y_toy, 2))
+ _compute_risk_precision(
+ lambdas, y_preds_proba, np.expand_dims(y_toy, 2)
+ )
  with pytest.raises(ValueError, match=r".*could not be broadcast*"):
  _compute_risk_precision(lambdas, y_preds_proba, y_toy[:-1])
 
@@ -188,4 +196,4 @@ def test_invalid_shape_alpha_hb() -> None:
 def test_delta_none_ltt(delta: Optional[float]) -> None:
  """Test error message when invalid delta"""
  with pytest.raises(ValueError, match=r".*Invalid delta"):
- _ltt_procedure(r_hat, alpha, delta, n)
+ _ltt_procedure(r_hat, alpha, delta, n)

mapie/tests/test_multi_label_classification.py

@@ -324,7 +324,7 @@ def test_results_for_partial_fit(strategy: str) -> None:
 
 @pytest.mark.parametrize("strategy", [*STRATEGIES])
 @pytest.mark.parametrize(
- "alpha", [np.array([0.3, 0.4]), [0.3, 0.4], (0.3, 0.4)]
+ "alpha", [np.array([0.05, 0.1]), [0.05, 0.1], (0.05, 0.1)]
 )
 def test_results_for_alpha_as_float_and_arraylike(
  strategy: str, alpha: Any