add matching_indices

README.md

@@ -131,6 +131,7 @@ At minium, the metric requires the original time-series and the generated time-s
 - **return_distance**: (bool, optional): Whether to return the distance matrix. Default is False.
 - **return_matching**: (bool, optional): Whether to return the matching matrix. Default is False.
 - **return_each_features**: (bool, optional): Whether to return the results for each feature. Default is False.
+- **return_matching_indices**: (bool, optional): Whether to return the indices of matching. Default is False.
 - **return_coverages**: (bool, optional): Whether to return the coverages. Default is False.
 - **return_all**: (bool, optional): Whether to return all the results. Default is False.
 - **dtype**: (str, optional): The data type used for computation. Default is "float32".
@@ -154,6 +155,10 @@ Let prediction instances be $P = \{p_1, p_2, \ldots, p_n\}$ and reference instan
 - **distance**: (numpy.ndarray): The distance matrix between the generated instances and the reference instances.
 - **match**: (numpy.ndarray): The matching matrix between the generated instances and the reference instances.
 - **match_inv**: (numpy.ndarray): The matching matrix between the reference instances and the generated instances.
+- **g2r_index**: (numpy.ndarray): The index of the reference instance that matches the generated instance for each feature.
+- **r2g_index**: (numpy.ndarray): The index of the generated instance that matches the reference instance for each feature.
+- **mean_g2r_index**: (numpy.ndarray): The index of the reference instance that matches the generated instance.
+- **mean_r2g_index**: (numpy.ndarray): The index of the generated instance that matches the reference instance.
 
 <!-- #### Values from Popular Papers -->
 <!-- *Give examples, preferrably with links to leaderboards or publications, to papers that have reported this metric, along with the values they have reported.* -->

__main__.py

@@ -43,7 +43,7 @@ results = metric.compute(
     predictions=predictions,
     references=references,
     batch_size=args.batch_size,
-    num_processes=args.num_process,
+    num_processes=args.num_processes,
     return_each_features=True,
     return_coverages=True,
     dtype=args.dtype,

matching_series.py

@@ -229,6 +229,7 @@ class matching_series(evaluate.Metric):
         return_matching: bool = False,
         return_each_features: bool = False,
         return_coverages: bool = False,
+        return_matching_indices: bool = False,
         return_all: bool = False,
         dtype=np.float32,
         eps: float = 1e-8,
@@ -249,6 +250,7 @@ class matching_series(evaluate.Metric):
             return_matching: bool, optional: Whether to return the matching matrix. Default is False.
             return_each_features: bool, optional: Whether to return the results for each feature. Default is False.
             return_coverages: bool, optional: Whether to return the coverages. Default is False.
+            return_matching_indices: bool, optional: Whether to return the indices of matching. Default is False.
             return_all: bool, optional: Whether to return all the results. Default is False.
             dtype: str, optional: The data type used for computation. Default is "float32".
             eps: float, optional: The epsilon value to avoid division by zero. Default is 1e-8.
@@ -301,8 +303,9 @@ class matching_series(evaluate.Metric):
             dtype=dtype,
         )
 
+        distance_mean = distance.mean(axis=-1)
         metrics = self._compute_metrics(
-            distance=distance.mean(axis=-1),
+            distance=distance_mean,
             eps=eps,
             cuc_n_calculation=cuc_n_calculation,
             cuc_n_samples=cuc_n_samples,
@@ -341,6 +344,19 @@ class matching_series(evaluate.Metric):
                         "coverages_features": [m["coverages"] for m in metrics_feature],
                     }
                 )
+        if return_matching_indices:
+            # matching index
+            g2r_index = distance.argmin(axis=1)
+            r2g_index = distance.argmin(axis=0)
+            # mean
+            mean_g2r_index = distance_mean.argmin(axis=1)
+            mean_r2g_index = distance_mean.argmin(axis=0)
+
+            out["g2r_index"] = g2r_index
+            out["r2g_index"] = r2g_index
+            out["mean_g2r_index"] = mean_g2r_index
+            out["mean_r2g_index"] = mean_r2g_index
+
         return out
 
     def compute_cuc(