feat(advisor): Authority Router — Advisor collection-agnostisch, KB-2026.1-Gewinn im Produktpfad

Der Advisor fan-outete bisher selbst ueber eine feste Liste expliziter Collections (advisor-rag.ts) und umging damit das #61-Scope-Routing (das nur den Default-Pfad routet) → der gemessene +28-Retrieval-Gewinn (CB-100: 53→81, 0 Regr) kam nie beim Antwort-LLM an. Dieser Router zieht den Fan-out in die Retriever-Schicht: - SDK: LegalRAGClient.Retrieve() + POST /sdk/v1/rag/retrieve {query, top_k} — fan-outet server-seitig ueber die Broad-Authority-Base + die KB-2026.1-Slice bei inKBScope, merge+dedup, sortiert nach Authority-Score (rerankByAuthority je Collection), top-K. Index-Warmup vor dem nebenlaeufigen Fan-out (Map-Race-frei). Per-Env via RAG_ROUTER_COLLECTIONS. - admin: advisor-rag.ts ruft EINMAL /retrieve statt 6-fach expliziter Collections. Advisor ist collection-agnostisch (Vertrag Compiler→Collections→Retriever→Advisor); COMPLIANCE_COLLECTIONS/searchCollection entfernt. Validierung: Go-Unit (Router-Selektion, dedup); e2e gegen dev-Qdrant (echter Retrieve(), CB-100-Stichprobe stride 5): OLD-hit 11/20 → NEW-hit 15/20, GAIN 4 (alle DS-Guidance), REGR 0 — reproduziert den +28/0-Regr durch den Produktionscode. TS-Tests auf den Single-/retrieve-Call angepasst. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-06-30 14:13:09 +02:00
parent af11d21f6e
commit 1e5aaf7103
7 changed files with 386 additions and 51 deletions
@@ -82,6 +82,43 @@ func (h *RAGHandlers) Search(c *gin.Context) {
 	})
 }

+// RetrieveRequest is the Authority Router request: a query only, no collection — the router decides
+// which collections to query (broad authority base + the in-scope KB-2026.1 slice).
+type RetrieveRequest struct {
+	Query string `json:"query" binding:"required"`
+	TopK  int    `json:"top_k,omitempty"`
+}
+
+// Retrieve is the Authority Router endpoint. The Advisor calls this with ONLY a query and stays
+// collection-agnostic; the router fans out over the authority base + the in-scope slice, merges by
+// authority score, and returns the unified top-K. Response shape matches Search (query/results/
+// count/assessment) so existing consumers parse it unchanged.
+// POST /sdk/v1/rag/retrieve
+func (h *RAGHandlers) Retrieve(c *gin.Context) {
+	var req RetrieveRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+
+	if req.TopK <= 0 || req.TopK > 20 {
+		req.TopK = 8
+	}
+
+	results, err := h.ragClient.Retrieve(c.Request.Context(), req.Query, req.TopK)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": "RAG retrieve failed: " + err.Error()})
+		return
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"query":      req.Query,
+		"results":    results,
+		"count":      len(results),
+		"assessment": ucca.Assess(results),
+	})
+}
+
 // ListRegulations returns the list of available regulations in the corpus.
 // GET /sdk/v1/rag/regulations
 func (h *RAGHandlers) ListRegulations(c *gin.Context) {