small refactor

contentious_contexts.py

@@ -97,20 +97,21 @@ class ContentiousContexts(datasets.GeneratorBasedBuilder):
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
-                gen_kwargs={"filepaths": [ann_file, text_file], "split": "train",},
+                gen_kwargs={"ann_file": ann_file, "text_file": text_file},
             ),
         ]
 
-    def _generate_examples(self, filepaths, split):
-        annotations = pd.read_csv(filepaths[0], dtype="object")
-        texts = pd.read_csv(filepaths[1], dtype="object")
+    def _generate_examples(self, ann_file, text_file):
+        annotations = pd.read_csv(open(ann_file), dtype="object")
+        texts = pd.read_csv(open(text_file), dtype="object")
         annotations.fillna("", inplace=True)
         texts.fillna("", inplace=True)
         for _, row in texts.iterrows():
-            data_point = {}
-            data_point["extract_id"] = row["extract_id"]
-            data_point["target"] = row["target_compound_bolded"]
-            data_point["text"] = row["text"]
+            data_point = {
+                "extract_id": row["extract_id"],
+                "target": row["target_compound_bolded"],
+                "text": row["text"],
+            }
             annotator_responses = annotations[
                 annotations["extract_id"] == row["extract_id"]
             ]