Allow use of ORJSONParser

[PaarthShah]

Fixes #144

Tested manually by setting the orjson render/parser as the default in settings.py: not sure if there's another way to be doing that/including it in the tests

[PaarthShah]

@vbabiy Any objections to this?

[piraka9011]

FWIW this library doesn't seem maintained, but in case someone comes across this PR/issue in the future, this PR is a step in the right direction but, CamelCaseJSONParser.parse also needs to be updated to:

def parse(self, stream, media_type=None, parser_context=None):
    try:
        data = super().parse(stream, media_type, parser_context)
        return underscoreize(data, **self.json_underscoreize)
    except ValueError as exc:
        raise ParseError("JSON parse error - %s" % str(exc))

Otherwise, as you'll notice in the current implementation, it always uses json.loads which negates the whole point behind improving the performance of deserialization.