Fix the license term for the dataset (#25)

* Update README.md

* Add README to bench/data

* fix

* fix

README.md

@@ -208,6 +208,8 @@ Licensed under either of
 
 at your option.
 
+For softwares under `bench/data`, follow the license terms of each software.
+
 ## Contribution
 
 Unless you explicitly state otherwise, any contribution intentionally submitted

bench/data/README.md

@@ -0,0 +1,8 @@
+# Datasets for benchmarking
+
+These datasets are copied from third party repositories.
+
+* `unidic`: [National Institute for Japanese Language and Linguistics](https://ccd.ninjal.ac.jp/unidic/)
+* `sherlock.txt`: [Project Gutenberg](https://www.gutenberg.org/ebooks/1661)
+* `wagahaiwa_nekodearu.txt`: [Aozora Bunko](https://www.aozora.gr.jp/cards/000148/card789.html)
+* `words_100000`: [fst crate](https://github.com/BurntSushi/fst/blob/master/data/words-100000)