Improve (better arguments validation, avoiding repeated creation of validator objects, etc) _validate_obj_json() in metadata.py and supporting funcs

[candleindark]

This PR make the following improvement relating to validating metadata instances against a DANDI model's JSON schema.

1. Redefine the interface of _validate_obj_json() so that it take a jsonschema validator instead of a JSON schema as argument to avoid recreation of validators for the same schema.
2. Define a utility function, `jsonschema_validator(), for creating an appropriate JSON schema validator given a schema.
   1. Ensure validity of the schema is checked to prevent undefined behavior. (A bug fix)
   2. Utilized the "$schema" property in a provide schema to determine the appropriate validator class
3. Define a utility function, validate_json(), to validate a data instance in general, without the missing_ok parameter.
4. Define a utility function, dandi_jsonschema_validator, that decides the default/fallback validator class to used based on the "schemaVersion" property of the given schema instead of the "schemaVersion" field of a given instance. (A bug fix)
5. Define utility functions, _get_jsonschema_validator()and _get_jsonschema_validator_local(), with cache to avoid repeated creation of jsonschema validators for the same schemas. Note: functools.cache() instead of functools.lru_cache() is used, for improved performance, in these functions because valid inputs to these functions have been limited by validation, so there is no worry about the cache growing without bound.
6. Reimplement _validate_obj_json() using the aforementioned utility functions
7. Rename schema_map in metadata.py to SCHEMA_MAP for the reference dictionary is a constant.
8. Add extensive tests to functions covered in this PR.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.24%. Comparing base (87cc55f) to head (e32f32d).
Report is 8 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #278      +/-   ##
==========================================
- Coverage   97.63%   93.24%   -4.40%     
==========================================
  Files          16       16              
  Lines        1775     1983     +208     
==========================================
+ Hits         1733     1849     +116     
- Misses         42      134      +92     

Flag	Coverage Δ
unittests	93.24% <100.00%> (-4.40%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[yarikoptic]

wouldn't that be called for the same schema over and over again for every instance?

[candleindark]

@yarikoptic I think the code is good now. However, I did quite a bit of validation of function args. I could have used @validate_call from Pydantic to remove some code duplications and boilerplates. Let me know if you want me to do that as well, but I think I spent enough time on this PR already.