update

UTS_Text.py

@@ -12,9 +12,29 @@ _CITATION = """\
 """
 
 _BASE_URL = "https://huggingface.co/datasets/undertheseanlp/UTS_Text_v1/resolve/main/data/"
-TRAIN_FILE = "UTS_Text_v1.txt"
+TRAIN_FILE = "train.txt"
+VALIDATION_FILE = "validation.txt"
+TEST_FILE = "test.txt"
+
+
+class UTSTextConfig(datasets.BuilderConfig):
+    """BuilderConfig"""
+
+    def __init__(self, **kwargs):
+        super(UTSTextConfig, self).__init__(**kwargs)
 
 class UTSText(datasets.GeneratorBasedBuilder):
+    """UTS Word Tokenize datasets"""
+    VERSION = datasets.Version("1.0.0")
+    BUILDER_CONFIGS = [
+        UTSTextConfig(
+            name="small", version=VERSION, description="UTS_Text Small"),
+        UTSTextConfig(
+            name="base", version=VERSION, description="UTS_Text Base"),
+        UTSTextConfig(
+            name="large", version=VERSION, description="UTS_Text Large")
+    ]
+
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
@@ -30,18 +50,28 @@ class UTSText(datasets.GeneratorBasedBuilder):
     
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
-        DATA_PATH = os.path.join(_BASE_URL, self.config.name)
-        train_file = dl_manager.download(os.path.join(DATA_PATH, "train.txt"))
-        # dev_file = dl_manager.download(os.path.join(DATA_PATH, "dev.txt"))
-        # test_file = dl_manager.download(os.path.join(DATA_PATH, "test.txt"))
-        data_files = {
-            "train": train_file,
-            # "dev": dev_file,
-            # "test": test_file,
-        }
+        subset_folder = self.config.name
+        train_file = dl_manager.download(os.path.join(_BASE_URL, subset_folder, TRAIN_FILE))
+        validation_file = dl_manager.download(os.path.join(_BASE_URL, subset_folder, VALIDATION_FILE))
+        test_file = dl_manager.download(os.path.join(_BASE_URL, subset_folder, TEST_FILE))
+
         splits = [
-            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": data_files["train"]}),
-            # datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": data_files["dev"]}),
-            # datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": data_files["test"]}),
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"filepath": train_file}
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"filepath": validation_file}
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"filepath": test_file}
+            ),
         ]
         return splits
+
+    def _generate_examples(self, filepath):
+        with open(filepath, encoding="utf-8") as f:
+            for line in f:
+                yield line.strip()