- Updated V3 CRUD routers to return 400 Bad Request for database schema
errors (unknown columns) across all list and search endpoints.
- Fixed serialization issue in nested patch endpoint.
- Overhauled Section 7 of Frontend Integration Guide to document HTTP
status code mappings for common error categories.
- Updated api_crud_v3 and api_crud_v3_nested to detect 'database_schema'
errors (like Unknown Column) and return a 400 Bad Request instead of
a generic 500 Internal Server Error.
- Added missing error handling for sql_select failure in get_child_obj_li.
- Modify sql_select to return False on database exceptions instead of an empty result
- Update V3 Search endpoint to catch False results and return a 500 Internal Server Error
- Enhances error visibility for frontend developers and prevents misleading 200 OK responses during SQL errors
- Re-apply safe guest auth and passcode-to-JWT endpoint
- Consolidate AccountContext with token_payload and role flags
- Restore documentation for new guest flows and public read whitelists
- Fix 403 error in get_obj_li by allowing optional account context
- Patched request_jwt to strip privileged IDs when signing with public keys
- Updated AccountContext and V3 dependencies to preserve JWT payloads for guests
- Whitelisted Archive, Post, Event, and other core objects for public read access
- Added 'default_qry_str' to Event searchable fields
- Added test_e2e_jwt_guest_auth.py for security verification
- Added 'x-ae-ignore-extra-fields' header to support stripping unknown fields in POST/PATCH.
- Added automatic resolution of '*_id_random' strings to integer IDs in 'sanitize_payload'.
- Refactored 'post_obj' to return structured (field -> message) validation errors in 'meta.details'.
- Updated 'mk_resp' to support non-string 'details' in response metadata.
- Added 'tests/verify_feedback_fixes.py' to validate logic changes.
Ref: V3 API Refinement Feedback from mcp_agent.
- Updated app/db_sql.py to capture SQL exceptions in thread-local storage for later retrieval.
- Implemented format_db_error() in app/lib_api_crud_v3.py to clean up raw MariaDB error strings.
- Added POST /v3/crud/{obj_type}/validate endpoint for dry-run payload validation.
- Updated main and nested routers to bubble up validation and database errors into the response 'meta.details' field.
- Added tests/test_v3_error_bubbling.py to verify formatting logic.
- Created app/lib_schema_v3.py to isolate database and Pydantic model introspection.
- Updated app/routers/api_crud_v3.py to use get_object_schema_info(), completing the modularization.
- Finalized refactoring plan documentation in documentation/REFACTOR_API_CRUD_V3.md.
- Created app/routers/api_crud_v3_nested.py to handle all parent-child relational routes.
- Updated app/routers/api_crud_v3.py to include the nested router, significantly reducing file size.
- Documented Phase 2 completion in documentation/REFACTOR_API_CRUD_V3.md.
- Created app/lib_api_crud_v3.py to house core security, filtering, and sanitization logic.
- Implemented reusable sanitize_payload() to generically strip virtual lookup fields (*_id_random) and view-only fields (fields_to_exclude_from_db).
- Updated app/routers/api_crud_v3.py to use the new library and consolidated sanitization across all Create/Update endpoints.
- Documented Phase 1 completion in documentation/REFACTOR_API_CRUD_V3.md.
- Added comprehensive docstrings to api_crud_v3.py explaining multi-tenancy, sanitization, and soft-delete logic.
- Finalized Address and Contact models/mappings with correct validators and field maps.
- Consolidated test suite in tests/ directory.
- Added 'external_person_id' to Post searchable fields.
- Updated api_crud_v3.py to respect 'fields_to_exclude_from_db' model attribute.
- Cleaned up old verification scripts (moved to tests/).
- Added Pydantic validators to Person_Base to handle null values for given_name and allow_auth_key, ensuring database NOT NULL constraints are met.
- Updated api_crud_v3.py (POST and PATCH) to filter out virtual *_id_random fields from data payloads before database operations to prevent "Unknown column" errors.
- Updated GEMINI.md with session progress.
- Implemented 'Isolation Mode' in api_crud_v3.py to bypass circular dependencies.
- Locally defined AccountContext and auth dependencies to ensure stable boot.
- Added site_domain lookup exception for guest users in search endpoint.
- Maintained agent_bridge disablement in main.py for stability.
- Enhanced AuthContext with role-aware fields (administrator, manager, super).
- Implemented deferred database lookups for user roles in get_v3_auth_context.
- Added global account isolation in api_crud_v3.py using check_account_access and apply_forced_account_filter.
- Hardened all V3 CRUD endpoints (GET, POST, PATCH, DELETE) and nested routes with ownership verification.
- Enforced forced account filtering at the SQL level for Listing and Searching.
- Updated documentation with details on the new security and data isolation architecture.
