|
|
71d32d11ee
|
feat(iit): V3.2 data consistency + project isolation + admin config redesign + Chinese labels
Summary:
- Refactor timeline API to read from qc_field_status (SSOT) instead of qc_logs
- Add field-issues paginated API with severity/dimension/recordId filters
- Add LEFT JOIN field_metadata + qc_event_status for Chinese display names
- Implement per-project ChatOrchestrator cache and SessionMemory isolation
- Redesign admin IIT config tabs (REDCap -> Fields -> KB -> Rules -> Members)
- Add AI-powered QC rule generation (D3 programmatic + D1/D5/D6 LLM-based)
- Add clickable warning/critical detail Modal in ReportsPage
- Auto-dispatch eQuery after batch QC via DailyQcOrchestrator
- Update module status documentation to v3.2
Backend changes:
- iitQcCockpitController: rewrite getTimeline from qc_field_status, add getFieldIssues
- iitQcCockpitRoutes: add field-issues route
- ChatOrchestrator: per-projectId cached instances
- SessionMemory: keyed by userId::projectId
- WechatCallbackController: resolve projectId from iitUserMapping
- iitRuleSuggestionService: dimension-based suggest + generateD3Rules
- iitBatchController: call DailyQcOrchestrator after batch QC
Frontend changes:
- AiStreamPage: adapt to new timeline structure with dimension tags
- ReportsPage: clickable stats cards with issue detail Modal
- IitProjectDetailPage: reorder tabs, add AI rule generation UI
- iitProjectApi: add TimelineIssue, FieldIssueItem types and APIs
Status: TypeScript compilation verified, no new lint errors
Made-with: Cursor
|
2026-03-02 14:29:59 +08:00 |
|
|
|
7c3cc12b2e
|
feat(iit): Complete CRA Agent V3.0 P1 - ChatOrchestrator with LLM Function Calling
P1 Architecture: Lightweight ReAct (Function Calling loop, max 3 rounds)
Core changes:
- Add ToolDefinition/ToolCall types to LLM adapters (DeepSeek + CloseAI + Claude)
- Replace 6 old tools with 4 semantic tools: read_report, look_up_data, check_quality, search_knowledge
- Create ChatOrchestrator (~160 lines) replacing ChatService (1,442 lines)
- Wire WechatCallbackController to ChatOrchestrator, deprecate ChatService
- Fix nullable content (string | null) across 12+ LLM consumer files
E2E test results: 8/8 scenarios passed (100%)
- QC report query, critical issues, patient data, trend, on-demand QC
- Knowledge base search, project overview, data modification refusal
Net code reduction: ~1,100 lines
Tested: E2E P1 chat test 8/8 passed with DeepSeek API
Made-with: Cursor
|
2026-02-26 14:27:09 +08:00 |
|
