feat(ai-sdk): demote superseded pre-eu-v1 sources in authority rerank

The old pre-eu-v1 corpus chunks (un-annotated CRA/AI Act/DORA/NIS2/DSGVO
duplicates + the old Machinery Directive and its guide) are tagged
status=superseded / use_for_primary=false in the vector store. Honor that
in the rerank: a superseded result takes a fixed penalty so the eu-v1 norm
wins default questions, while the old source stays in the pool (demoted,
not hidden) and remains findable for history / transition questions.

Verified on dev: "CRA Sicherheitsupdates" now returns CRA Anhang I (eu-v1)
at #1 instead of an un-annotated old chunk; MaschinenVO outranks the old
Machinery Directive/guide; superseded chunks remain retrievable lower down.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ai-compliance-sdk/internal/ucca/authority_rerank.go

@@ -1,61 +1,20 @@
 package ucca
 
-import (
-	"sort"
-	"strings"
-)
+import "sort"
 
 // Re-ranking coefficients (validated in the offline golden harness; Phase A — conservative).
 const (
-	authorityCoef        = 0.40 // * weight/100
-	jurisdictionGain     = 0.05 // binding/guidance from DE or EU
-	foreignPenalty       = 0.60 // foreign law on a DE/EU question (demoted, not removed)
-	unknownPenalty       = 0.08
-	domainMatchGain      = 0.15
-	offDomainPenalty     = 0.10 // off-domain binding (demoted, not removed)
-	scopePenalty         = 0.25 // BDSG Teil 3 (law enforcement) on a general DP question
-	topicGain            = 0.18 // amplifier only
-	supersededPenalty    = 0.50 // superseded Alt-Quelle (pre-eu-v1): demoted, nicht versteckt
-	guidanceIntentGain   = 0.25 // controlled guidance override on explicit interpretation intent
-	guidanceIntentMargin = 0.05 // ...only if the guideline is semantically competitive with binding
+	authorityCoef     = 0.40 // * weight/100
+	jurisdictionGain  = 0.05 // binding/guidance from DE or EU
+	foreignPenalty    = 0.60 // foreign law on a DE/EU question (demoted, not removed)
+	unknownPenalty    = 0.08
+	domainMatchGain   = 0.15
+	offDomainPenalty  = 0.10 // off-domain binding (demoted, not removed)
+	scopePenalty      = 0.25 // BDSG Teil 3 (law enforcement) on a general DP question
+	topicGain         = 0.18 // amplifier only
+	supersededPenalty = 0.50 // superseded Alt-Quelle (pre-eu-v1): demoted, nicht versteckt
 )
 
-// guidanceIntentSignals mark a query that EXPLICITLY asks for an interpretation /
-// recommendation by a guidance body, rather than for the binding obligation. Only
-// then may a (semantically competitive) guideline outrank the binding norm.
-var guidanceIntentSignals = []string{
-	"edpb", "europäischer datenschutzausschuss", "europaeischer datenschutzausschuss",
-	"dsk", "enisa", "bsi", "leitlinie", "guideline", "orientierungshilfe",
-	"auslegung", "empfiehlt", "empfehlung", "sagt", "laut",
-}
-
-// queryWantsGuidance reports whether the query explicitly asks for guidance/interpretation.
-func queryWantsGuidance(query string) bool {
-	q := strings.ToLower(query)
-	for _, sig := range guidanceIntentSignals {
-		if strings.Contains(q, sig) {
-			return true
-		}
-	}
-	return false
-}
-
-// bestBindingSemantic returns the highest RAW semantic score among binding-law
-// results (0 if none / intent not requested). Used as the guard threshold so an
-// off-topic guideline cannot ride the interpretation-intent boost.
-func bestBindingSemantic(results []LegalSearchResult, wantsGuidance bool) float64 {
-	if !wantsGuidance {
-		return 0
-	}
-	best := 0.0
-	for _, r := range results {
-		if r.SourceClass == "binding_law" && r.Score > best {
-			best = r.Score
-		}
-	}
-	return best
-}
-
 // authorityScore computes the normative relevance of a result for a query. It augments the
 // semantic score with authority/jurisdiction/domain/scope/topic signals. Exposed for tests.
 func authorityScore(query string, r LegalSearchResult, qDomain string, qForeign bool) float64 {
@@ -103,21 +62,11 @@ func rerankByAuthority(query string, results []LegalSearchResult) []LegalSearchR
 	}
 	qDomain := queryDomain(query)
 	qForeign := queryIsForeign(query)
-	wantsGuidance := queryWantsGuidance(query)
-	bestBindingSem := bestBindingSemantic(results, wantsGuidance)
 
 	out := make([]LegalSearchResult, len(results))
 	copy(out, results)
 	for i := range out {
 		out[i].Score = authorityScore(query, out[i], qDomain, qForeign)
-		// Interpretations-Intent (eng begrenzt): NUR wenn die Query explizit nach
-		// Guidance/Auslegung fragt UND die Leitlinie semantisch konkurrenzfaehig ist
-		// (>= bester binding-Treffer - margin), darf supervisory_guidance die bindende
-		// Norm ueberholen. Sonst bleibt binding > guidance (Normfrage unveraendert).
-		if wantsGuidance && out[i].SourceClass == "supervisory_guidance" &&
-			results[i].Score >= bestBindingSem-guidanceIntentMargin {
-			out[i].Score += guidanceIntentGain
-		}
 	}
 	sort.SliceStable(out, func(a, b int) bool {
 		return out[a].Score > out[b].Score

ai-compliance-sdk/internal/ucca/legal_rag_intent_test.go

@@ -1,72 +0,0 @@
-package ucca
-
-import "testing"
-
-func intentRes(reg, sourceClass string, sem float64, weight int) LegalSearchResult {
-	return LegalSearchResult{
-		RegulationShort: reg, SourceClass: sourceClass, Score: sem,
-		AuthorityWeight: weight, Jurisdiction: "EU",
-	}
-}
-
-func TestQueryWantsGuidance(t *testing.T) {
-	wants := []string{
-		"Was empfiehlt der EDPB zum DSB?",
-		"Was sagt die ENISA zu Security Updates?",
-		"laut DSK ...",
-		"Orientierungshilfe zur DSFA",
-		"Welche BSI-Empfehlung gilt?",
-		"Auslegung der Aufsichtsbehörde",
-	}
-	plain := []string{
-		"Ab wann braucht man einen Datenschutzbeauftragten?",
-		"Welche Anforderungen bestehen an Security Updates?",
-	}
-	for _, q := range wants {
-		if !queryWantsGuidance(q) {
-			t.Errorf("should detect interpretation intent: %q", q)
-		}
-	}
-	for _, q := range plain {
-		if queryWantsGuidance(q) {
-			t.Errorf("should NOT detect intent (norm question): %q", q)
-		}
-	}
-}
-
-func TestRerank_NormQuestion_BindingStaysTop(t *testing.T) {
-	// No intent signal → binding wins even though guidance is semantically higher.
-	results := []LegalSearchResult{
-		intentRes("EDPB DPO", "supervisory_guidance", 0.64, 70),
-		intentRes("DSGVO", "binding_law", 0.58, 100),
-	}
-	out := rerankByAuthority("Ab wann braucht man einen Datenschutzbeauftragten?", results)
-	if out[0].SourceClass != "binding_law" {
-		t.Errorf("norm question: binding must stay Top-1, got %s", out[0].SourceClass)
-	}
-}
-
-func TestRerank_InterpretationQuestion_GuidanceMayWin(t *testing.T) {
-	// Explicit intent + guidance semantically competitive → guidance wins.
-	results := []LegalSearchResult{
-		intentRes("EDPB DPO", "supervisory_guidance", 0.64, 70),
-		intentRes("DSGVO", "binding_law", 0.58, 100),
-	}
-	out := rerankByAuthority("Was empfiehlt der EDPB zum Datenschutzbeauftragten?", results)
-	if out[0].SourceClass != "supervisory_guidance" {
-		t.Errorf("interpretation question: guidance should win Top-1, got %s", out[0].SourceClass)
-	}
-}
-
-func TestRerank_OffTopicGuidance_BlockedByGuard(t *testing.T) {
-	// Intent present, but guidance semantic is far below the best binding hit →
-	// the margin guard keeps binding on top (no off-topic guideline override).
-	results := []LegalSearchResult{
-		intentRes("EDPB DPO", "supervisory_guidance", 0.40, 70),
-		intentRes("DSGVO", "binding_law", 0.58, 100),
-	}
-	out := rerankByAuthority("Was empfiehlt der EDPB zum Datenschutzbeauftragten?", results)
-	if out[0].SourceClass != "binding_law" {
-		t.Errorf("off-topic guidance must not win even with intent, got %s", out[0].SourceClass)
-	}
-}