Feature: add to_dict() and to_json() to PyJWK and PyJWKSet

[pranavkumaarofficial]

Summary

PyJWK and PyJWKSet can be constructed from dictionaries via from_dict(),
but there is no public way to serialize them back. The only workaround is
accessing the private _jwk_data attribute or using Algorithm.to_jwk(),
which loses metadata like kid, use, and alg.

This adds to_dict() and to_json() methods to both classes, enabling
lossless round-tripping of JWK objects while preserving all metadata fields.

Changes

• jwt/api_jwk.py: Added to_dict() and to_json() to PyJWK and PyJWKSet
• tests/test_api_jwk.py: Added 10 tests covering round-trips for RSA (public + private), EC, HMAC, OKP key types, metadata preservation, JSON serialization, keyset round-trips, and unusable key filtering
• CHANGELOG.rst: Added entry under [Unreleased] > Added

Design notes

PyJWK.to_dict() returns a shallow copy of the internal _jwk_data
dictionary that was passed in during construction. This is intentional:
it preserves all original JWK fields (including kid, use, alg,
key_ops) that Algorithm.to_jwk() would drop.

PyJWKSet.to_dict() returns {"keys": [...]} containing only the keys
that were successfully parsed during construction. Keys with unsupported
algorithms that were skipped on input will not appear in the output.

How to test

pytest tests/test_api_jwk.py -v -k "to_dict or to_json"

Closes #1106