Switch simplejson to orjson

[neyberson]

orjson simplifies and solves some issues:

• the datetime issue https://github.com/ijl/orjson?tab=readme-ov-file#datetime
• the options about numpy https://github.com/ijl/orjson?tab=readme-ov-file#numpy
• use utf-8 "has strict UTF-8 conformance, more correct than the standard library" https://github.com/ijl/orjson?tab=readme-ov-file#orjson

So, just adding the option OPT_SERIALIZE_NUMPY in the orjson.dumps(), avoid the issues with numpy and using orjson simplifies the code

[ayjayt]

👍 on my list for today

[ayjayt]

@neyberson this isn't passing tests

Problem with {"id":2,"result":{}}{bye}
 in json: unexpected content after document: line 1 column 21 (char 20)

This message is correct but the new parser is reading it as one message. It's a json message followed by a control message. This needs to be resolved.

[ayjayt]

@neyberson the problem is that with package json also we have it set up to receive partial json messages and also json messages that have junk on the end- it still pulls out the whole message correctly, thats why its read_jsons and not read_json

[ayjayt]

failing tests

[ayjayt]

also @neyberson this is particularly an issue with windows which, on all other machines, the OS seems to fill and flush the buffer message per message but windows messages sometimes come in partially, sometimes grouped, etc