Update arxiv_pii.py

arxiv_pii.py

@@ -38,11 +38,11 @@ class ArXivPII(datasets.GeneratorBasedBuilder):
         example_id = 0
         df_pages = pd.read_csv(pages_path, index_col="arxiv_id")
         df_labels = pd.read_csv(labels_path, index_col="arxiv_id")
-        for document, d in df_pages.iterrows():
+        for arxiv_id, page in df_pages.iterrows():
             labels = []
-            if document in df_labels.index:
-                rows = df_labels.loc[[document]]
+            if arxiv_id in df_labels.index:
+                rows = df_labels.loc[[arxiv_id]]
                 for _, e in rows.iterrows():
                     labels.append({"start": e.start, "end": e.end, "label": e.label})
-            yield example_id, {"text": d.text, "labels": labels}
+            yield example_id, {"text": page.text, "labels": labels}
             example_id += 1