Add support for non-rust implemented tokenization for __getitem__ method. (#24039)

* Add support for non-rust implemented tokenization for `__getitem__` method.

* Update for error message on adding new sub-branch for `__item__` method.

---------

Co-authored-by: liuyang17 <liuyang17@zhihu.com>

src/transformers/tokenization_utils_base.py

@@ -233,15 +233,20 @@ class BatchEncoding(UserDict):
         etc.).
 
         If the key is an integer, get the `tokenizers.Encoding` for batch item with index `key`.
+
+        If the key is a slice, returns the value of the dict associated to `key` ('input_ids', 'attention_mask', etc.)
+        with the constraint of slice.
         """
         if isinstance(item, str):
             return self.data[item]
         elif self._encodings is not None:
             return self._encodings[item]
+        elif isinstance(item, slice):
+            return {key: self.data[key][slice] for key in self.data.keys()}
         else:
             raise KeyError(
-                "Indexing with integers (to access backend Encoding for a given batch index) "
+                "Invalid key. Only three types of key are available: "
-                "is not available when using Python based tokenizers"
+                "(1) string, (2) integers for backend Encoding, and (3) slices for data subsetting."
             )
 
     def __getattr__(self, item: str):