Change default parameters for Shapley estimator

This should significantly speed up some calculations for larger graphs. In this regard, also increased the default samples for the anomaly attribution to balance the reduced number of Shapley run.

Signed-off-by: Patrick Bloebaum <bloebp@amazon.com>

dowhy/gcm/anomaly.py

@@ -94,7 +94,7 @@ def attribute_anomalies(
     anomaly_samples: pd.DataFrame,
     anomaly_scorer: Optional[AnomalyScorer] = None,
     attribute_mean_deviation: bool = False,
-    num_distribution_samples: int = 1500,
+    num_distribution_samples: int = 3000,
     shapley_config: Optional[ShapleyConfig] = None,
 ) -> Dict[Any, np.ndarray]:
     """Estimates the contributions of upstream nodes to the anomaly score of the target_node for each sample in

dowhy/gcm/shapley.py

@@ -44,7 +44,7 @@ class ShapleyConfig:
     def __init__(
         self,
         approximation_method: ShapleyApproximationMethods = ShapleyApproximationMethods.AUTO,
-        num_permutations: int = 2000,
+        num_permutations: int = 25,
         num_subset_samples: int = 5000,
         min_percentage_change_threshold: float = 0.05,
         n_jobs: Optional[int] = None,
@@ -110,7 +110,7 @@ def estimate_shapley_values(
         if num_players <= 5:
             approximation_method = ShapleyApproximationMethods.EXACT
         else:
-            approximation_method = ShapleyApproximationMethods.EARLY_STOPPING
+            approximation_method = ShapleyApproximationMethods.PERMUTATION
 
     if approximation_method == ShapleyApproximationMethods.EXACT:
         return _estimate_shapley_values_exact(set_func=set_func, num_players=num_players, n_jobs=shapley_config.n_jobs)
@@ -369,7 +369,7 @@ def _approximate_shapley_values_via_early_stopping(
                         permutation, evaluated_subsets, full_subset_result, empty_subset_result
                     )
 
-            if run_counter > max_num_permutations:
+            if run_counter * num_permutations_per_run > max_num_permutations:
                 break
 
             new_shap_proxy = np.array(shapley_values)
@@ -579,12 +579,15 @@ def parallel_job(input_subset: Tuple[int], parallel_random_seed: int) -> Union[f
 
         return set_func(np.array(input_subset))
 
+    if isinstance(evaluation_subsets, set):
+        evaluation_subsets = list(evaluation_subsets)
+
     random_seeds = np.random.randint(np.iinfo(np.int32).max, size=len(evaluation_subsets))
     subset_results = parallel_context(
-        delayed(parallel_job)(subset_to_evaluate, int(random_seed))
-        for subset_to_evaluate, random_seed in tqdm(
-            zip(evaluation_subsets, random_seeds),
-            desc="Evaluate set function",
+        delayed(parallel_job)(evaluation_subsets[i], int(random_seeds[i]))
+        for i in tqdm(
+            range(len(evaluation_subsets)),
+            desc="Evaluating set functions...",
             position=0,
             leave=True,
             disable=not config.show_progress_bars or not show_progressbar,