update data loader

xCodeEval.py

@@ -6,6 +6,7 @@ import zipfile
 from collections import defaultdict
 import datasets
 import textwrap
+from multiprocessing import Pool, cpu_count
 logger = datasets.logging.get_logger(__name__)
 
 _CITATION = """\
@@ -185,8 +186,8 @@ _CUSTOM_FEATURES = {
 	"insert_cnt": datasets.Value("int32"),
 	"fix_ops_cnt": datasets.Value("int32"),
 	"difficulty": datasets.Value("int32"),
-	"prob_desc_sample_inputs": datasets.Sequence(datasets.Value("string")),
-	"prob_desc_sample_outputs": datasets.Sequence(datasets.Value("string")),
+	# "prob_desc_sample_inputs": datasets.Sequence(datasets.Value("string")),
+	# "prob_desc_sample_outputs": datasets.Sequence(datasets.Value("string")),
 }
 
 _PROB_DESC_TEXT_FEATURES = {
@@ -1958,12 +1959,13 @@ class xCodeEval(datasets.GeneratorBasedBuilder):
               		for feature in self.config.text_features
                 }
 		if self.config.name in _UNIT_TEST_REQ_TASK:
-			features["hidden_unit_test"] = datasets.features.Sequence(
-                        {
-                            "input": datasets.Value("string"),
-                            "output": datasets.Sequence(datasets.Value("string")),
-                        }
-                    ),
+			features["hidden_unit_test"] = datasets.Value("string")
+		# 	features["hidden_unit_test"] = datasets.features.Sequence(
+        #                 {
+        #                     "input": datasets.Value("string"),
+        #                     "output": datasets.Sequence(datasets.Value("string")),
+        #                 }
+        #             ),
 		return datasets.DatasetInfo(
 			# This is the description that will appear on the datasets page.
 			description=self.config.description + "\n" + _DESCRIPTION,
@@ -2038,7 +2040,7 @@ class xCodeEval(datasets.GeneratorBasedBuilder):
 	 			}
 			),
 		]
-  
+		
 	def _generate_examples(self, filepaths, problem_description_file = None, unit_test_db_file = None):
 		"""This function returns the examples"""
 		task_name = self.config.name
@@ -2060,17 +2062,30 @@ class xCodeEval(datasets.GeneratorBasedBuilder):
 			with open(filepath) as rp:
 				for idx, line in enumerate(rp):
 					sample = json.loads(line)
-					if task_name not in _PROBLEM_DESC_REQ_TASK and problem_descriptions is not None:
+					for pre_feature in list(sample.keys()):
+						if pre_feature not in _TEXT_FEATURES[task_name]:
+							sample.pop(pre_feature)
+					for feature in _TEXT_FEATURES[task_name]:
+						if feature not in sample:
+							sample[feature] = ""
+					if task_name in _UNIT_TEST_REQ_TASK:
+						sample["hidden_unit_test"] = ""
+					if task_name not in _PROBLEM_DESC_REQ_TASK or problem_descriptions is None:
 						yield idx, sample # if problem_description_file is None then unit_test_db_file should be None
+						continue
 					
 					# if problem_description_file is not None, the sample has `src_uid`
 					src_uid = sample["src_uid"]
 					prob_desc = problem_descriptions[src_uid]
 					for key, ckey in _PROB_DESC_TEXT_FEATURES.items():
-						sample[ckey] = prob_desc[key]
+						if ckey == "prob_desc_sample_inputs" or ckey == "prob_desc_sample_outputs":
+							sample[ckey] = json.dumps(prob_desc[key])
+						else:
+							sample[ckey] = prob_desc[key]
 
-					if task_name not in _UNIT_TEST_REQ_TASK and unit_test_db is not None:
+					if task_name not in _UNIT_TEST_REQ_TASK or unit_test_db is None:
 						yield idx, sample
-    
-					sample["hidden_unit_test"] = unit_test_db[src_uid]
+						continue
+					
+					sample["hidden_unit_test"] = json.dumps(unit_test_db[src_uid])
 					yield idx, sample