Convert custom user_data to token extension format for Japanese tokenizer

[adrianeboyd]

Description

Convert the user_data values so that they can be loaded as custom token extensions for inflection, reading_form, sub_tokens, and lemma.

Types of change

Checklist

• I have submitted the spaCy Contributor Agreement.
• I ran the tests, and all new and existing tests passed.
• My changes don't require a change to the documentation, or if they do, I've added all required information.

[Techno-coder]

Was there a reason why this PR was reverted? I can't seem to find any discussions on it 😅. I would have thought accessing inflections by Token would be more natural.

[adrianeboyd]

Yes, I agree that it seems more natural, but we have a policy that the core spacy library never sets custom extensions so that there's no chance for it to introduce conflicts with users' custom extensions, which can potentially have the same names.