[Breaking] Feature: include '々' iteration character in kanji, exclude from JA punctuation

[DJTB]

Closes #113

Currently, Wanakana considers 々 to be punctuation (due to where it lies in unicode ranges, alongside other pure punctuation symbols), though in practice it functions like a kanji character. This causes minor issues such as tokenize() splitting up words like 人々 incorrectly and isKanji() reporting false for 人々.

This PR now considers 々 to be a kanji character, and excludes it from punctuation checks.
Please see modified tests to see which functions are affected and #113 for prior discussion, as well as DJTB/react-furi#6.

This doesn't really affect the core usage of Wanakana (as an IME) but peripheral utils that we provide. Whenever 々 comes up currently it is behaving incorrectly, so despite considering this a fix I don't want consumers updating if they see a patch bump (5.1.1). Since it changes existing behaviour perhaps it should be a major release?

@mimshwright @scottnicolson @vietqhoang do you have any concerns merging this?
Do you prefer a major 6.0.0 or minor 5.2.0 with breaking changes listed in release/changelog?

[DJTB]

Also merged tokenize docs update since this modifies them as well.
Fixes #156 and closes #157

[vietqhoang]

Did a first pass review. Nothing really stood out to me as far as the changes go. My first thoughts is I do agree the repeater should be considered more of a kanji than a punctuation.

As for the preference on how to increment the version, to me this is more of a minor version than a major version. The changes don't really add anything significant nor does it break the API.

[DJTB]

As for the preference on how to increment the version, to me this is more of a minor version than a major version. The changes don't really add anything significant nor does it break the API.

You're right that a major bump doesn't make much sense, I was being overly cautious since it changes some behaviour 😅(albeit for the better).