update script to remove fewshot split

data/fewshot-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5a9616244730b16c2206f28abda97ceb484d79381a33c474380d3f4369a25402
-size 19101

data/test-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d7ffb343ba1006a2d87ff746e7974eaa80848a3cb82b234defe1e24875db3d1f
-size 1019953

enem_challenge_disabled.py

@@ -243,34 +243,24 @@ class ENEMChallenge(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager):
         downloaded_file = dl_manager.download(FILE_URL)
-        fewshot = _get_train_examples()
+        explanation_questions = {d['id']:d['explanation'] for d in _get_train_examples()}
         dataset = {
             '2022': datasets.load_dataset(DATASET_PATH, name="2022", split="train"),
             '2023': datasets.load_dataset(DATASET_PATH, name="2023", split="train")
         }
 
-        #filter fewshot
-        to_filter = [int(d['id'].split('_')[1]) for d in fewshot]
-        dataset['2022'] = dataset['2022'].filter(lambda x: int(x['id'].split('_')[1]) not in to_filter)
-
         return [
             datasets.SplitGenerator(
-                name=datasets.Split.TEST,
+                name=datasets.Split.TRAIN,
                 gen_kwargs={
                     "data_list": [downloaded_file, dataset],
-                    "types": ["zip", "dataset"]
-                }
-            ),
-            datasets.SplitGenerator(
-                name='fewshot',
-                gen_kwargs={
-                    "data_list": [fewshot],
-                    "types": ["list"]
+                    "types": ["zip", "dataset"],
+                    "explanation_questions": explanation_questions
                 }
-            )        
+            )     
         ]
 
-    def _generate_examples(self, data_list, types):
+    def _generate_examples(self, data_list, types, explanation_questions):
         for data, type in zip(data_list, types):
             if type == "zip":
                 with zipfile.ZipFile(data, 'r') as zip_ref:
@@ -301,26 +291,8 @@ class ENEMChallenge(datasets.GeneratorBasedBuilder):
                                 "question": document["context"] + '\n' + document["question"],
                                 "choices": choices,
                                 "answerKey": document["label"].upper(),
-                                "explanation": None,
+                                "explanation": explanation_questions.get(doc_id, None)
                             }
-            if type == "list":
-                for document in data:
-                    choices = {
-                        "text": document["options"],
-                        "label": ["A", "B", "C", "D", "E"]
-                    }
-                    doc_id = document['id']
-                    yield doc_id, {
-                        "id": doc_id,
-                        "question_number": int(doc_id.split('_')[1]),
-                        "exam_id": document['exam'],
-                        "exam_year": document['exam'].split('_')[0],
-                        "nullified": None,
-                        "question": document["context"] + '\n' + document["question"],
-                        "choices": choices,
-                        "answerKey": document["label"].upper(),
-                        "explanation": document["explanation"],
-                    }
             if type == "dataset":
                 for exam_id in data:
                     dataset = data[exam_id]
@@ -342,6 +314,6 @@ class ENEMChallenge(datasets.GeneratorBasedBuilder):
                             "question": document["question"],
                             "choices": choices,
                             "answerKey": document["label"].upper(),
-                            "explanation": None,
+                            "explanation": explanation_questions.get(doc_id, None)
                         }