iterate trought image names and store it in a list for each dataset

NIH-Chest-X-ray-dataset.py

@@ -91,11 +91,42 @@ class XChest(datasets.GeneratorBasedBuilder):
 
 
 	def _split_generators(self, dl_manager):
-		data_files = dl_manager.download_and_extract(_URLS['image_urls'])
+		# Get the image names that belong to the train-val dataset
 		train_val_list = get(_URLS['train_val_list']).iter_lines()
 		train_val_list = set([x.decode('UTF8') for x in train_val_list])
-		print(train_val_list)
-		pass
+		print(train_val_list))
+
+		# Create list for store the name of the images for each dataset
+		train_files = []
+		test_files = []
+		
+		# Download batches
+		data_files = dl_manager.download_and_extract(_URLS['image_urls'])
+
+		# Iterate trought image folder and check if they belong to
+		# the trainset or testset
+		for batch in data_files:
+		  for img in batch:
+		    if img in train_val_list:
+			train_files.append(img)
+		    else:
+			test_files.append(img)
+
+		print(train_files)
+		print(test_files)
+		
+		return [
+		    datatsets.SplitGenerator(
+			name=datasets.Split.TRAIN,
+			gen_kwargs={
+				'files': dl_manager.iter_files()
+			}
+
+		    ),
+		    datasets.SplitGenerator(
+
+		    )
+		]
 
 	def _generate_examples(self, files):
 		pass