Snap

clustering_files.py

@@ -100,12 +100,10 @@ class ClusteringFiles(datasets.GeneratorBasedBuilder):
                 "end": row['end'],
                 "readable_start": row['readable_start'],
                 "readable_end": row['readable_end'],
-                "vp": row["vp"]  # Keep as numpy array
+                "vp": row["vp"]
             })
 
         for idx, file_data in enumerate(files.values()):
-            yield idx, file_data
-
-# Ensure numpy arrays are properly handled by datasets library
-datasets.config.NUMPY_ARRAYS_AVAILABLE = True
-datasets.config.NUMPY_DTYPE_TO_PYARROW_DTYPE[np.dtype("float32")] = datasets.Value("float32")
+            for segment in file_data["segments"]:
+                segment["vp"] = segment["vp"].tolist()
+            yield idx, file_data