Add per-tool in-memory call counters to the MCP module and a
GET /api/v1/status/mcp endpoint that joins them with the live tool
registry and endpoint config, then render it as a new card on the
System Status page with a one-click client-config copy button.
- app/mcp/stats.py: lock-guarded MCPStatsTracker (the mcp SDK runs sync
tool bodies via anyio.to_thread.run_sync, so this is genuinely
multi-threaded, unlike the async REST routes)
- app/mcp/server.py: instrument search_regulations, add get_mcp_status()
- app/config/settings.py: optional MCP_PUBLIC_URL override, required
because the Vite proxy and reverse proxies rewrite the Host header
- StatusPage.tsx: MCP Server card, joins the existing parallel fetch
Counters are process-local by design; token usage is already persisted
by ModelUsageTracker since MCP calls route through ask().
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- New backend/app/mcp/ module: MCPServer instance with a single
search_regulations tool backed by the existing AgentConversationService.
- MCPAuthMiddleware reuses existing JWT auth (no new auth mechanism).
- Mounted at /mcp/ in api/main.py via Streamable HTTP transport; wired the
MCP session manager into the existing lifespan() via AsyncExitStack
(app.mount() does not propagate nested ASGI lifespans automatically).
- Fixed a doubled /mcp/mcp path by setting streamable_http_path to "/"
(MCPServer.streamable_http_app() defaults to registering its own /mcp route).
- Verified end-to-end with the real mcp Python client: list_tools() returns
search_regulations, auth correctly 401s without or with an invalid token.
- 7 new tests, 76 total (up from 69), all passing.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Removed multiple failed document entries from `documents.json`.
- Added a new document entry with updated metadata and changed the index name to `regulations_dense_1024_v2`.
- Updated architecture documentation to reflect changes in the Milvus collection name.
- Adjusted requirements by removing the sqlalchemy dependency.
- Modified test cases to align with new document structure and naming conventions.
- Introduced a new test file for Milvus vector index runtime recovery and error handling.
- Updated assertions in various test files to ensure compatibility with the new schema.
feat(agent): update import for agent session service
feat(openai): add context truncation check in OpenAI answer generator
docs(README): update frontend environment file conventions
fix(vite): default local frontend development to local backend
Move session-related responsibilities into a new application-layer AgentSessionService (and AgentSessionFeedbackResult dataclass), provide a bootstrap factory (get_agent_session_service), and update agent API routes to call the service instead of accessing ConversationStore directly. Routes now translate ValueError into 404 responses and use service methods for get/list/history/delete/feedback. Also update package exports and docs/READMEs to declare the backend architecture authority, enforce api -> application -> domain ports -> infrastructure boundaries, and call out legacy services/workflows as migration-only. These changes centralize session logic in the application layer and tighten architecture guidance for future backend work.
- Updated LocalDocumentParser to include raw_layouts and artifact_prefix from settings.
- Added new documents with failure reasons and metadata to documents.json for better error tracking.
- Created a new documentation file detailing the Aliyun ingest implementation process.
- Updated RFC to reflect changes in the parsing backend and embedding dimensions.
- Modified tests to accommodate the new embedding dimension of 1024 and updated parser and chunk builder assertions.
- Verified migration configurations to ensure correct settings for embedding model and backend.
wangwei