[camden11]

Why

The Python SDK has no client for the OAuth v3 token endpoints, so Python users can't move off the v1 token API. This adds HubSpot().oauth.v3.advanced_api with get_token, introspect_token, and revoke_token. It matches the Node client in HubSpot/hubspot-api-nodejs#702.

The second commit regenerates the client from an updated spec. The spec now declares no response body for revoke_token, so its return type changes from Error to None. In Python only the docstring :rtype changes; behavior is the same.

Notes for reviewers

• Generated with openapi-generator python-legacy 6.5.0 and formatted with black 24.10.0 -l 200. Use that exact black version: black 26 changes docstring style.
• Known gap: python-legacy merges each oneOf union into one class, which keeps only the last member's enum and required fields. With client_side_validation=True, every response raises ValueError on token_use. Client-side validation is off by default, as in the rest of the SDK.
• client_credentials is untested: the test app gets BAD_SCOPES.

Testing

• tests/spec/oauth/v3/test_v3.py covers discovery, plus deserialization for get_token and introspect_token. All of tests/spec passes (105/105).
• QA against a test portal, run before the regeneration commit:
  • get_token with authorization_code and with refresh_token
  • a CRM call using the new token
  • introspection of access and refresh tokens
  • a bad token raising ApiException 400
  • revoke_token returning None, after which introspection reports active: false
• A QA re-run on the regenerated client is in progress.

🤖 Generated with Claude Code