Update semeval2017.py

semeval2017.py

@@ -35,10 +35,13 @@ _LICENSE = ""
 # TODO: Add link to the official dataset URLs here
 
 _URLS = {
-    "test": "test.jsonl"
+    "test": "test.jsonl",
+    "train": "train.jsonl",
+    "valid": "valid.jsonl"
 }
 
 
+
 # TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
 class SemEval2017(datasets.GeneratorBasedBuilder):
     """TODO: Short description of my dataset."""
@@ -108,6 +111,14 @@ class SemEval2017(datasets.GeneratorBasedBuilder):
 
         data_dir = dl_manager.download_and_extract(_URLS)
         return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={
+                    "filepath": data_dir['train'],
+                    "split": "train",
+                },
+            ),
             datasets.SplitGenerator(
                 name=datasets.Split.TEST,
                 # These kwargs will be passed to _generate_examples
@@ -116,6 +127,14 @@ class SemEval2017(datasets.GeneratorBasedBuilder):
                     "split": "test"
                 },
             ),
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={
+                    "filepath": data_dir['valid'],
+                    "split": "valid",
+                },
+            ),
         ]
 
     # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`