Fix to issue #1.

[sangwinc]

https://www.compart.com/en/unicode/category/Pe
https://www.compart.com/en/unicode/category/Ps

[sangwinc]

@christianp Do you accept this, and are there any further additions to this project? I'd like to keep STACK in sync with this list!

[christianp]

Sorry, I completely forgot about this! I'm off work this week for half term, but I'll look at it when I'm back.

[christianp]

@sangwinc now I've looked at #1 again, I see I didn't add these to symbols.json. Instead, I added a file brackets.json, listing bracket characters that aren't normalised by NFKD to the standard parenthesis, curly brace, or square bracket characters. The only ones in this PR that weren't in my list were the light left/right tortoise shell ornaments, so I've added those.

For convenience, I'll repeat what I said in my last comment to #1: you can detect bracket characters by looking for character classes Ps and Po, normalise with NFKD, then apply the mapping from brackets.json if necessary.