Fix WordCount example.

The UnicodeScriptTokenizer was failing because the input Tensor had rank 0 (i.e. scalar), and the API requires rank 1.

PiperOrigin-RevId: 268715414

tensorflow_federated/python/research/analytics/BUILD

+package(default_visibility = ["//tensorflow_federated/python/research:__subpackages__"])
+
+licenses(["notice"])  # Apache 2.0
+
+exports_files(
+    srcs = ["wordcount.ipynb"],
+)