update generator.

gaokao_bench.py

@@ -34,7 +34,7 @@ _HOMEPAGE = "https://cevalbenchmark.com"
 
 _LICENSE = "Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License"
 
-_URL = r"https://huggingface.co/datasets/AsakusaRinne/Gaokao_bench/tree/main/Multiple-choice_Questions"
+_URL = r"https://huggingface.co/datasets/AsakusaRinne/gaokao_bench/blob/main/gaokao_bench.zip"
 
 task_list = [
     '2010-2013_English_MCQs',
@@ -87,35 +87,35 @@ class GaokaoBench(datasets.GeneratorBasedBuilder):
             citation=_CITATION,
         )
 
-    # def _split_generators(self, dl_manager):
-    #     data_dir = dl_manager.download_and_extract(_URL)
-    #     task_name = self.config.name
-    #     return [
-    #         datasets.SplitGenerator(
-    #             name=datasets.Split.TEST,
-    #             gen_kwargs={
-    #                 "filepath": os.path.join(
-    #                     data_dir, "test", f"{task_name}_test.csv"
-    #                 ),
-    #             },
-    #         ),
-    #         datasets.SplitGenerator(
-    #             name=datasets.Split("val"),
-    #             gen_kwargs={
-    #                 "filepath": os.path.join(
-    #                     data_dir,  "val", f"{task_name}_val.csv"
-    #                 ),
-    #             },
-    #         ),
-    #         datasets.SplitGenerator(
-    #             name=datasets.Split("dev"),
-    #             gen_kwargs={
-    #                 "filepath": os.path.join(
-    #                     data_dir, "dev", f"{task_name}_dev.csv"
-    #                 ),
-    #             },
-    #         ),
-    #     ]
+    def _split_generators(self, dl_manager):
+        data_dir = dl_manager.download_and_extract(_URL)
+        task_name = self.config.name
+        return [
+            # datasets.SplitGenerator(
+            #     name=datasets.Split.TEST,
+            #     gen_kwargs={
+            #         "filepath": os.path.join(
+            #             data_dir, "test", f"{task_name}_test.csv"
+            #         ),
+            #     },
+            # ),
+            # datasets.SplitGenerator(
+            #     name=datasets.Split("val"),
+            #     gen_kwargs={
+            #         "filepath": os.path.join(
+            #             data_dir,  "val", f"{task_name}_val.csv"
+            #         ),
+            #     },
+            # ),
+            datasets.SplitGenerator(
+                name=datasets.Split("MCQ"),
+                gen_kwargs={
+                    "filepath": os.path.join(
+                        data_dir, "Multiple-choice_Questions", f"{task_name}.csv"
+                    ),
+                },
+            ),
+        ]
 
     def _generate_examples(self, filepath):
         df = pd.read_csv(filepath,encoding="utf-8")