Fix test cases.

src/pydvl/value/shapley/truncated.py

@@ -255,11 +255,7 @@ def truncated_montecarlo_shapley(
     # including the ones that are running
     n_submitted_jobs = 2 * n_jobs
 
-    accumulated_result = ValuationResult.zeros(
-        algorithm=algorithm,
-        indices=u.data.indices,
-        data_names=u.data.data_names,
-    )
+    accumulated_result = ValuationResult.zeros(algorithm=algorithm)
 
     with init_executor(max_workers=n_jobs, config=config) as executor:
         futures = set()

tests/value/shapley/test_classwise.py

@@ -1,6 +1,8 @@
 """
 Test cases for the class wise shapley value.
 """
+import random
+from random import seed
 from typing import Dict, Tuple
 
 import numpy as np
@@ -239,7 +241,7 @@ def linear_classifier_cs_scorer_args_exact_solution_use_default_score() -> Tuple
                 ]
             )
         ),
-        {"rtol": 0.15},
+        {"atol": 0.05},
     )
 
 
@@ -263,7 +265,7 @@ def linear_classifier_cs_scorer_args_exact_solution_use_default_score_norm() ->
                 ]
             )
         ),
-        {"rtol": 0.15},
+        {"atol": 0.05},
     )
 
 
@@ -738,8 +740,13 @@ def linear_classifier_cs_scorer_args_exact_solution_use_add_idx_empty_set() -> T
     )
 
 
-@pytest.mark.parametrize("n_samples", [500])
-@pytest.mark.parametrize("n_resample_complement_sets", [3])
+@pytest.mark.parametrize("n_samples", [500], ids=lambda x: "n_samples={}".format(x))
+@pytest.mark.parametrize(
+    "n_resample_complement_sets",
+    [20],
+    ids=lambda x: "n_resample_complement_sets={}".format(x),
+)
+@pytest.mark.parametrize("seed", [13, 42, 666], ids=lambda x: "seed={}".format(x))
 @pytest.mark.parametrize(
     "linear_classifier_cs_scorer_args_exact_solution",
     [
@@ -754,8 +761,12 @@ def test_classwise_shapley(
     linear_classifier_cs_scorer_args_exact_solution: Tuple[Dict, ValuationResult],
     n_samples: int,
     n_resample_complement_sets: int,
+    seed: int,
     request,
 ):
+    np.random.seed(seed)
+    random.seed(seed)
+
     args, exact_solution, check_args = request.getfixturevalue(
         linear_classifier_cs_scorer_args_exact_solution
     )