- API tokens now verify project belongs to token owner before access
- Researcher tokens only access research/general docs in owner's projects
- Developer tokens only access development/general docs in owner's projects
- Viewer tokens have read-only access to all doc types in owner's projects
- Add test for cross-user project access prevention
- Add api_tokens table with role-based access (researcher, developer, viewer)
- Add POST /auth/token/generate endpoint for creating tokens
- Add GET /auth/tokens endpoint for listing user's tokens
- Add DELETE /auth/tokens/{token_id} endpoint for revoking tokens
- Add agent_type field to documents (research, development, general)
- Implement role-based access control for documents:
- researcher: access to research and general documents
- developer: access to development and general documents
- viewer: read-only access
- Update document model and schemas with agent_type field
- Add comprehensive tests for API token functionality
- All existing tests pass (73 total)
- Add DISABLE_REGISTRATION setting (default False) in app/config.py
- Return 403 when registration is disabled in POST /auth/register
- Add test_register_disabled test
- Added PATCH endpoint for projects (supports name and/or description)
- Added test_patch_project test case
- Verify folder_id already supported in POST /api/v1/projects/{project_id}/documents
- Add outgoing_links (JSON) and backlinks_count to Document model
- POST /documents/{id}/detect-links — detect [[uuid]] patterns in content
- GET /documents/{id}/backlinks — documents referencing this doc
- GET /documents/{id}/outgoing-links — documents this doc references
- GET /documents/{id}/links — combined incoming + outgoing
- GET /projects/{id}/graph — full project relationship graph
- GET /search/quick — fuzzy search (Quick Switcher Cmd+K)
- GET /projects/{id}/documents/search — project-scoped search
- GET /documents/{id}/export — markdown|json export
- GET /projects/{id}/export — json|zip export
- 27 new tests
- Script migrates documents where content (markdown) exists but tiptap_content is NULL
- Idempotent: only processes documents needing migration
- Reports count of processed documents and any errors
- Includes post-migration validation
- Extend Document model with reasoning_type, confidence, reasoning_steps, model_source, tiptap_content fields
- Add new endpoints:
- GET /documents/{id}/reasoning - Get reasoning metadata
- PATCH /documents/{id}/reasoning - Update reasoning metadata
- GET /documents/{id}/reasoning-panel - Get reasoning panel data for UI
- POST /documents/{id}/reasoning-steps - Add reasoning step
- DELETE /documents/{id}/reasoning-steps/{step} - Delete reasoning step
- GET /documents/{id}/content?format=tiptap|markdown - Get content in TipTap or Markdown
- PUT /documents/{id}/content - Update content (supports both TipTap JSON and Markdown)
- Add TipTap to Markdown conversion
- Update database schema with new columns
- Add comprehensive tests for all new endpoints
- All 37 tests passing
