fix: formatted

embeddings-benchmark · Jun 2, 2024 · 8d3fc1b · 8d3fc1b
1 parent 53c477f
commit 8d3fc1b
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 7 deletions.
diff --git a/mteb/evaluation/evaluators/RerankingEvaluator.py b/mteb/evaluation/evaluators/RerankingEvaluator.py
@@ -283,7 +283,7 @@ def nAUC_scores(
         """Computes normalized Area Under the Curve on a set of evaluated instances as presented in the paper https://arxiv.org/abs/2402.12997
 
         Args:
-            all_conf_scores: Confidence scores for all instances, with length `len(samples)` 
+            all_conf_scores: Confidence scores for all instances, with length `len(samples)`
             metrics: Metric scores for all instances, with length `len(samples)`
             metric_name: Name of the metric (mrr or ap)
 

diff --git a/mteb/evaluation/evaluators/utils.py b/mteb/evaluation/evaluators/utils.py
@@ -307,13 +307,13 @@ def confidence_scores(sim_scores: List[float]) -> Dict[str, float]:
             - `diff1`: Difference between highest and second highest similarity scores
     """
     sim_scores_sorted = sorted(sim_scores)[::-1]
-    
+
     cs_max = sim_scores_sorted[0]
     cs_std = np.std(sim_scores)
     if len(sim_scores) > 1:
         cs_diff1 = sim_scores_sorted[0] - sim_scores_sorted[1]
     elif len(sim_scores) == 1:
-        cs_diff1 = 0.
+        cs_diff1 = 0.0
 
     conf_scores = {"max": cs_max, "std": cs_std, "diff1": cs_diff1}
 
@@ -323,7 +323,7 @@ def confidence_scores(sim_scores: List[float]) -> Dict[str, float]:
 def nAUC(
     conf_scores: np.ndarray,
     metrics: np.ndarray,
-    abstention_rates: np.ndarray = np.linspace(0,1,11)[:-1],
+    abstention_rates: np.ndarray = np.linspace(0, 1, 11)[:-1],
 ) -> float:
     """Computes normalized Area Under the Curve on a set of evaluated instances as presented in the paper https://arxiv.org/abs/2402.12997
     1/ Computes the raw abstention curve, i.e., the average evaluation metric at different abstention rates determined by the confidence scores
@@ -344,7 +344,7 @@ def nAUC(
     def abstention_curve(
         conf_scores: np.ndarray,
         metrics: np.ndarray,
-        abstention_rates: np.ndarray = np.linspace(0,1,11)[:-1],
+        abstention_rates: np.ndarray = np.linspace(0, 1, 11)[:-1],
     ) -> np.ndarray:
         """Computes the raw abstention curve for a given set of evaluated instances and corresponding confidence scores
 
@@ -365,7 +365,7 @@ def abstention_curve(
         return abst_curve
 
     def oracle_curve(
-        metrics: np.ndaray, abstention_rates: np.ndarray = np.linspace(0,1,11)[:-1]
+        metrics: np.ndaray, abstention_rates: np.ndarray = np.linspace(0, 1, 11)[:-1]
     ) -> np.ndarray:
         """Computes the oracle curve for a given set of evaluated instances
 

diff --git a/tests/test_RerankingEvaluator.py b/tests/test_RerankingEvaluator.py
@@ -47,7 +47,9 @@ def test_nAUC(self):
             [0.5, 0.4, 0.5, 0.2],
         ]
 
-        ap_scores = [self.evaluator.ap_score(y, x) for x, y in zip(pred_scores, is_relevant)]
+        ap_scores = [
+            self.evaluator.ap_score(y, x) for x, y in zip(pred_scores, is_relevant)
+        ]
         conf_scores = [self.evaluator.conf_scores(x) for x in pred_scores]
         nauc_scores_map = self.evaluator.nAUC_scores(conf_scores, ap_scores, "map")