Add some more test cases for tokenization and ascii folding

[SiarheiFedartsou]

👋 I did some awesome work for the Pelias project and would love for everyone to have a look at it and provide feedback.

---

Here's the reason for this change 🚀

There are some questions in discussion of this PR #498, so I'd like to propose to extend a bit existing test harness to kind of document current state of the things. So current schema:

• tokenizes by whitespace, hyphen, slashes (but doesn't by "dash" character for example - I'd expect ICU tokenizer to work differently here btw 🤔 )
• normalizes thai digits to Arabic ones (I think we can extrapolate it to any other digits writing system)
• removes tonal marks in Thai script
• we never make digits "glued" to the end of word a separate token
• we don't tokenize Asian languages which don't use whitespaces properly

---

Here's what actually got changed 👏

• Added tests

---

Here's how others can test the changes 👀

Run tests :)