Lift suite run #004 left two unresolved tail issues: - Q6 ("Forklift loader") ↔ Q7 ("Hazmat warehouse, cold storage") swap recordings as warm top-1 because their embeddings are within 0.20 cosine of each other. Distance gate can't tell them apart. - Q9 + Q15 lose paraphrase recovery when qwen2.5 rephrases past the 0.20 threshold. Distance says "drift too far"; sometimes the drift is real (skip), sometimes the paraphrase is still on-domain (don't want to skip). Multi-coord run #008's judge re-rating proved the LLM can distinguish: Q3 crane case landed at distance 0.23 (looks tight) but rating 1 (irrelevant). The judge sees domain mismatch the embedder doesn't. This commit lifts that pattern into the matrix substrate. Shape B inject now optionally routes every candidate through a judge gate before the rank insert lands. Distance + judge BOTH have to approve. internal/matrix/playbook.go: - InjectPlaybookMisses signature gains a query string + an optional InjectGate. nil gate preserves pre-judge-gating behavior (current tests already pass with nil). - New InjectGate interface + InjectGateFunc adapter for tests and non-LLM callers. - Per-candidate gate.Approve(query, hit) call inserted between the dedup and the inject. Rejected candidates skip silently; injected count reflects post-gate decision. internal/matrix/judge.go (new, ~140 lines): - LLMJudgeGate calls an Ollama-shape /api/chat endpoint with the same 1-5 staffing-rubric prompt that worked in multi_coord run #008. fail-closed on HTTP/JSON errors (don't inject if judge can't speak — better miss than wrong-domain). - NewLLMJudgeGate returns nil when URL or Model is empty, matching InjectGate's nil-means-no-judge semantics. internal/matrix/retrieve.go: - SearchRequest gains JudgeURL, JudgeModel, JudgeMinRating fields. Run() builds an LLMJudgeGate when set; passes nil otherwise. Backward compatible — existing callers see no behavior change. Tests: - TestInjectPlaybookMisses_GateRejectsCandidate (rejectAll → 0 injected, even with tight distance) - TestInjectPlaybookMisses_GateApprovesCandidate (approveAll → same as nil-gate behavior) - TestInjectPlaybookMisses_GateSeesCorrectQuery (gate receives CURRENT query + RECORDED query separately so it can score the (current, candidate) pair) - All 5 existing inject tests updated to new signature go test ./internal/matrix → all 8 inject tests pass. go test ./internal/matrix ./internal/shared ./cmd/{matrixd, queryd,pathwayd,observerd} → all green. STATE_OF_PLAY: - OPEN item #1 (judge-gated injection) closed. - DO NOT RELITIGATE adds the substrate-level judge-gate lock. - OPEN list now 5 rows (was 6). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
433 lines
14 KiB
Go
433 lines
14 KiB
Go
package matrix
|
|
|
|
import (
|
|
"encoding/json"
|
|
"testing"
|
|
)
|
|
|
|
func TestPlaybookEntry_Validate(t *testing.T) {
|
|
good := PlaybookEntry{
|
|
QueryText: "x", AnswerID: "y", AnswerCorpus: "z", Score: 0.5,
|
|
}
|
|
if err := good.Validate(); err != nil {
|
|
t.Errorf("good entry should validate: %v", err)
|
|
}
|
|
|
|
cases := []struct {
|
|
name string
|
|
entry PlaybookEntry
|
|
}{
|
|
{"empty query", PlaybookEntry{AnswerID: "y", AnswerCorpus: "z", Score: 0.5}},
|
|
{"empty answer id", PlaybookEntry{QueryText: "x", AnswerCorpus: "z", Score: 0.5}},
|
|
{"empty corpus", PlaybookEntry{QueryText: "x", AnswerID: "y", Score: 0.5}},
|
|
{"score too high", PlaybookEntry{QueryText: "x", AnswerID: "y", AnswerCorpus: "z", Score: 1.5}},
|
|
{"score negative", PlaybookEntry{QueryText: "x", AnswerID: "y", AnswerCorpus: "z", Score: -0.1}},
|
|
}
|
|
for _, c := range cases {
|
|
if err := c.entry.Validate(); err == nil {
|
|
t.Errorf("%s: expected validation error, got nil", c.name)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestPlaybookEntry_BoostFactor(t *testing.T) {
|
|
cases := []struct {
|
|
score float64
|
|
want float64
|
|
}{
|
|
{0.0, 1.0},
|
|
{0.5, 0.75},
|
|
{1.0, 0.5},
|
|
{-0.1, 1.0}, // clamped
|
|
{1.5, 0.5}, // clamped
|
|
}
|
|
for _, c := range cases {
|
|
got := PlaybookEntry{Score: c.score}.BoostFactor()
|
|
if abs(got-c.want) > 1e-9 {
|
|
t.Errorf("BoostFactor(score=%.2f): want %.4f, got %.4f", c.score, c.want, got)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestApplyPlaybookBoost_NoHitsLeaveResultsAlone(t *testing.T) {
|
|
results := []Result{
|
|
{ID: "a", Distance: 0.1, Corpus: "x"},
|
|
{ID: "b", Distance: 0.2, Corpus: "x"},
|
|
}
|
|
n := ApplyPlaybookBoost(results, nil)
|
|
if n != 0 {
|
|
t.Errorf("expected 0 boosted, got %d", n)
|
|
}
|
|
if results[0].ID != "a" || results[1].ID != "b" {
|
|
t.Errorf("results reordered without hits: %v", results)
|
|
}
|
|
}
|
|
|
|
func TestApplyPlaybookBoost_BoostMovesResultUp(t *testing.T) {
|
|
// Initial: a (0.10) beats b (0.20) beats c (0.30).
|
|
// Playbook says (answer=c, score=1.0) should be boosted → c's
|
|
// distance becomes 0.30 * 0.5 = 0.15. New ordering: a, c, b.
|
|
results := []Result{
|
|
{ID: "a", Distance: 0.10, Corpus: "x"},
|
|
{ID: "b", Distance: 0.20, Corpus: "x"},
|
|
{ID: "c", Distance: 0.30, Corpus: "x"},
|
|
}
|
|
hits := []PlaybookHit{
|
|
{PlaybookID: "p1", Distance: 0.05, Entry: PlaybookEntry{
|
|
AnswerID: "c", AnswerCorpus: "x", Score: 1.0,
|
|
}},
|
|
}
|
|
n := ApplyPlaybookBoost(results, hits)
|
|
if n != 1 {
|
|
t.Errorf("expected 1 boosted, got %d", n)
|
|
}
|
|
if results[0].ID != "a" || results[1].ID != "c" || results[2].ID != "b" {
|
|
t.Errorf("expected order a,c,b after boost; got %v", idsOf(results))
|
|
}
|
|
if abs(float64(results[1].Distance)-0.15) > 1e-6 {
|
|
t.Errorf("expected c distance 0.15 after boost; got %.4f", results[1].Distance)
|
|
}
|
|
}
|
|
|
|
func TestApplyPlaybookBoost_HighestScoreWinsForSameAnswer(t *testing.T) {
|
|
results := []Result{
|
|
{ID: "a", Distance: 0.30, Corpus: "x"},
|
|
}
|
|
// Two playbook hits both pointing at "a". Score=0.4 (weak boost)
|
|
// + Score=0.9 (strong boost). Strong should win — distance gets
|
|
// multiplied by 1-0.5*0.9 = 0.55, not by 1-0.5*0.4 = 0.80.
|
|
hits := []PlaybookHit{
|
|
{PlaybookID: "p_weak", Distance: 0.05, Entry: PlaybookEntry{
|
|
AnswerID: "a", AnswerCorpus: "x", Score: 0.4,
|
|
}},
|
|
{PlaybookID: "p_strong", Distance: 0.05, Entry: PlaybookEntry{
|
|
AnswerID: "a", AnswerCorpus: "x", Score: 0.9,
|
|
}},
|
|
}
|
|
ApplyPlaybookBoost(results, hits)
|
|
wantDist := 0.30 * 0.55
|
|
if abs(float64(results[0].Distance)-wantDist) > 1e-6 {
|
|
t.Errorf("strong-score boost should win: want %.4f, got %.4f", wantDist, results[0].Distance)
|
|
}
|
|
}
|
|
|
|
func TestApplyPlaybookBoost_CorpusAttributionRespected(t *testing.T) {
|
|
// Playbook references answer_id="a" in corpus="x".
|
|
// Results have answer_id="a" in corpus="y" — DIFFERENT corpus.
|
|
// Boost should NOT apply; the (id, corpus) tuple is the join key,
|
|
// not just id (otherwise different-corpus collisions would create
|
|
// false positives).
|
|
results := []Result{
|
|
{ID: "a", Distance: 0.30, Corpus: "y"},
|
|
}
|
|
hits := []PlaybookHit{
|
|
{PlaybookID: "p1", Distance: 0.05, Entry: PlaybookEntry{
|
|
AnswerID: "a", AnswerCorpus: "x", Score: 1.0,
|
|
}},
|
|
}
|
|
n := ApplyPlaybookBoost(results, hits)
|
|
if n != 0 {
|
|
t.Errorf("cross-corpus collision should not boost: got %d", n)
|
|
}
|
|
if abs(float64(results[0].Distance)-0.30) > 1e-6 {
|
|
// 1e-6 tolerance accounts for float32→float64 conversion;
|
|
// the assertion that matters is "unchanged from input."
|
|
t.Errorf("distance should be unchanged: got %.6f", results[0].Distance)
|
|
}
|
|
}
|
|
|
|
func TestPlaybookEntry_RoundTripJSON(t *testing.T) {
|
|
e := NewPlaybookEntry("forklift query", "w-12345", "workers", 0.85, []string{"chicago", "verified"})
|
|
raw, err := e.MarshalMetadata()
|
|
if err != nil {
|
|
t.Fatalf("marshal: %v", err)
|
|
}
|
|
got, err := UnmarshalPlaybookMetadata(raw)
|
|
if err != nil {
|
|
t.Fatalf("unmarshal: %v", err)
|
|
}
|
|
if got.QueryText != e.QueryText || got.AnswerID != e.AnswerID ||
|
|
got.AnswerCorpus != e.AnswerCorpus || got.Score != e.Score {
|
|
t.Errorf("round-trip mismatch: want %+v, got %+v", e, got)
|
|
}
|
|
if len(got.Tags) != 2 || got.Tags[0] != "chicago" {
|
|
t.Errorf("tags lost in round-trip: %v", got.Tags)
|
|
}
|
|
if got.RecordedAtNs == 0 {
|
|
t.Error("RecordedAtNs not set by NewPlaybookEntry")
|
|
}
|
|
}
|
|
|
|
func TestUnmarshalPlaybookMetadata_RejectsEmpty(t *testing.T) {
|
|
if _, err := UnmarshalPlaybookMetadata(json.RawMessage{}); err == nil {
|
|
t.Error("empty metadata should error")
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_AddsMissingAnswers locks Shape B's primary
|
|
// claim: when a playbook hit's answer isn't already in regular
|
|
// retrieval results, InjectPlaybookMisses appends a synthetic Result
|
|
// for it. Reality test playbook_lift_002 surfaced 0/2 paraphrase
|
|
// recoveries because the v0 boost-only stance couldn't promote
|
|
// answers that dropped out of the paraphrase's top-K.
|
|
func TestInjectPlaybookMisses_AddsMissingAnswers(t *testing.T) {
|
|
results := []Result{
|
|
{ID: "w-1", Corpus: "workers", Distance: 0.30},
|
|
{ID: "w-2", Corpus: "workers", Distance: 0.35},
|
|
}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-x",
|
|
Distance: 0.20, // current query is close to recorded query
|
|
Entry: PlaybookEntry{
|
|
QueryText: "recorded query",
|
|
AnswerID: "w-99", // NOT in results
|
|
AnswerCorpus: "workers",
|
|
Score: 1.0, // strong outcome → boost factor 0.5
|
|
},
|
|
},
|
|
}
|
|
out, injected := InjectPlaybookMisses("test query", results, hits, 0, nil)
|
|
if injected != 1 {
|
|
t.Fatalf("expected 1 injected, got %d", injected)
|
|
}
|
|
if len(out) != 3 {
|
|
t.Fatalf("expected len=3, got %d (%v)", len(out), idsOf(out))
|
|
}
|
|
// The injected result should be findable + carry the playbook
|
|
// provenance metadata flag.
|
|
var injectedResult *Result
|
|
for i := range out {
|
|
if out[i].ID == "w-99" {
|
|
injectedResult = &out[i]
|
|
break
|
|
}
|
|
}
|
|
if injectedResult == nil {
|
|
t.Fatal("w-99 not present in output")
|
|
}
|
|
// distance = 0.20 * 0.5 = 0.10 → near-top after caller re-sorts
|
|
if injectedResult.Distance < 0.099 || injectedResult.Distance > 0.101 {
|
|
t.Errorf("expected injected distance ~0.10, got %f", injectedResult.Distance)
|
|
}
|
|
var meta map[string]any
|
|
if err := json.Unmarshal(injectedResult.Metadata, &meta); err != nil {
|
|
t.Fatalf("decode meta: %v", err)
|
|
}
|
|
if v, _ := meta["playbook_injected"].(bool); !v {
|
|
t.Errorf("expected playbook_injected=true marker, got %v", meta)
|
|
}
|
|
if v, _ := meta["playbook_query_text"].(string); v != "recorded query" {
|
|
t.Errorf("expected recorded query in meta, got %v", v)
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_SkipsAnswersAlreadyPresent locks the
|
|
// boost-only-when-present property. If a playbook hit's answer is
|
|
// ALREADY in results, we don't duplicate-inject — ApplyPlaybookBoost
|
|
// has handled that case via in-place re-rank.
|
|
func TestInjectPlaybookMisses_SkipsAnswersAlreadyPresent(t *testing.T) {
|
|
results := []Result{
|
|
{ID: "w-1", Corpus: "workers", Distance: 0.30},
|
|
{ID: "w-99", Corpus: "workers", Distance: 0.40}, // ALREADY HERE
|
|
}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-x",
|
|
Distance: 0.20,
|
|
Entry: PlaybookEntry{
|
|
QueryText: "x", AnswerID: "w-99", AnswerCorpus: "workers", Score: 1.0,
|
|
},
|
|
},
|
|
}
|
|
out, injected := InjectPlaybookMisses("test query", results, hits, 0, nil)
|
|
if injected != 0 {
|
|
t.Errorf("expected 0 injected (answer already present), got %d", injected)
|
|
}
|
|
if len(out) != 2 {
|
|
t.Errorf("expected results unchanged at len=2, got %d", len(out))
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_DedupesPerAnswer locks: multiple playbook
|
|
// hits all pointing to the same missing answer collapse to ONE
|
|
// injection (the highest-scoring hit wins).
|
|
func TestInjectPlaybookMisses_DedupesPerAnswer(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-low",
|
|
Distance: 0.30,
|
|
Entry: PlaybookEntry{QueryText: "q1", AnswerID: "w-99", AnswerCorpus: "workers", Score: 0.4},
|
|
},
|
|
{
|
|
PlaybookID: "pb-high",
|
|
Distance: 0.30,
|
|
Entry: PlaybookEntry{QueryText: "q2", AnswerID: "w-99", AnswerCorpus: "workers", Score: 1.0},
|
|
},
|
|
}
|
|
out, injected := InjectPlaybookMisses("test query", results, hits, 0.5, nil) // explicit loose threshold so 0.30 hits qualify
|
|
if injected != 1 {
|
|
t.Errorf("expected 1 injection (deduped), got %d", injected)
|
|
}
|
|
// Score=1.0 (the high one) wins → boost factor 0.5 → distance 0.15
|
|
for _, r := range out {
|
|
if r.ID == "w-99" {
|
|
if r.Distance < 0.149 || r.Distance > 0.151 {
|
|
t.Errorf("expected distance from highest-score hit (~0.15), got %f", r.Distance)
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_GateRejectsCandidate locks the judge-gate
|
|
// path (OPEN item #1, closed by this commit). When the InjectGate
|
|
// returns false on a candidate, the candidate is skipped — even if
|
|
// distance would otherwise allow it. Closes the lift-suite tail
|
|
// issues where embedding said "tight" but a judge said "wrong domain."
|
|
func TestInjectPlaybookMisses_GateRejectsCandidate(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-x",
|
|
Distance: 0.10, // tight in cosine — would inject without gate
|
|
Entry: PlaybookEntry{
|
|
QueryText: "recorded crane operator query",
|
|
AnswerID: "w-99", AnswerCorpus: "workers", Score: 1.0,
|
|
},
|
|
},
|
|
}
|
|
rejectAll := InjectGateFunc(func(string, PlaybookHit) bool { return false })
|
|
out, injected := InjectPlaybookMisses("forklift loader query", results, hits, 0, rejectAll)
|
|
if injected != 0 {
|
|
t.Errorf("rejectAll gate should skip injection, got %d injected", injected)
|
|
}
|
|
if len(out) != 1 {
|
|
t.Errorf("results should be unchanged at len=1, got %d", len(out))
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_GateApprovesCandidate locks the
|
|
// always-approve gate path: behavior matches nil-gate (current
|
|
// distance-only filter). Useful for tests that want to assert
|
|
// "judge-gate API is wired" without an actual decision.
|
|
func TestInjectPlaybookMisses_GateApprovesCandidate(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-x",
|
|
Distance: 0.10,
|
|
Entry: PlaybookEntry{
|
|
QueryText: "x", AnswerID: "w-99", AnswerCorpus: "workers", Score: 1.0,
|
|
},
|
|
},
|
|
}
|
|
approveAll := InjectGateFunc(func(string, PlaybookHit) bool { return true })
|
|
out, injected := InjectPlaybookMisses("test query", results, hits, 0, approveAll)
|
|
if injected != 1 {
|
|
t.Errorf("approveAll gate should inject, got %d", injected)
|
|
}
|
|
if len(out) != 2 {
|
|
t.Errorf("results should grow to 2, got %d", len(out))
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_GateSeesCorrectQuery locks the gate's
|
|
// query+hit visibility — the gate must receive the CURRENT search's
|
|
// query (not the recorded one) so it can judge the (current_query,
|
|
// candidate) pair. The recorded query lives on hit.Entry.QueryText.
|
|
func TestInjectPlaybookMisses_GateSeesCorrectQuery(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "pb-x",
|
|
Distance: 0.10,
|
|
Entry: PlaybookEntry{
|
|
QueryText: "RECORDED",
|
|
AnswerID: "w-99", AnswerCorpus: "workers", Score: 1.0,
|
|
},
|
|
},
|
|
}
|
|
var seenQuery, seenRecordedQuery string
|
|
gate := InjectGateFunc(func(q string, h PlaybookHit) bool {
|
|
seenQuery = q
|
|
seenRecordedQuery = h.Entry.QueryText
|
|
return true
|
|
})
|
|
_, _ = InjectPlaybookMisses("CURRENT", results, hits, 0, gate)
|
|
if seenQuery != "CURRENT" {
|
|
t.Errorf("gate received query=%q, want CURRENT", seenQuery)
|
|
}
|
|
if seenRecordedQuery != "RECORDED" {
|
|
t.Errorf("gate received recorded=%q, want RECORDED", seenRecordedQuery)
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_RespectsInjectThreshold locks the
|
|
// cross-pollination defense added after run #003: hits whose playbook
|
|
// distance exceeds the inject threshold are skipped, preventing the
|
|
// "OSHA-30 forklift" recording from surfacing as warm top-1 for an
|
|
// unrelated dental-hygienist query just because their text vectors
|
|
// happened to fall within boost-threshold (0.5).
|
|
func TestInjectPlaybookMisses_RespectsInjectThreshold(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
// Two hits: one within tight inject threshold, one beyond it but
|
|
// within boost threshold. Only the tight one should inject.
|
|
hits := []PlaybookHit{
|
|
{
|
|
PlaybookID: "tight",
|
|
Distance: 0.10, // within inject (true paraphrase territory)
|
|
Entry: PlaybookEntry{QueryText: "q1", AnswerID: "w-tight", AnswerCorpus: "workers", Score: 1.0},
|
|
},
|
|
{
|
|
PlaybookID: "loose",
|
|
Distance: 0.40, // boost-eligible but inject-rejected
|
|
Entry: PlaybookEntry{QueryText: "q2", AnswerID: "w-loose", AnswerCorpus: "workers", Score: 1.0},
|
|
},
|
|
}
|
|
// Default threshold (0 → DefaultPlaybookMaxInjectDistance = 0.20)
|
|
out, injected := InjectPlaybookMisses("test query", results, hits, 0, nil)
|
|
if injected != 1 {
|
|
t.Errorf("expected 1 injection (only the tight hit qualifies), got %d", injected)
|
|
}
|
|
gotTight := false
|
|
for _, r := range out {
|
|
if r.ID == "w-tight" {
|
|
gotTight = true
|
|
}
|
|
if r.ID == "w-loose" {
|
|
t.Errorf("loose hit (distance > inject threshold) was injected anyway")
|
|
}
|
|
}
|
|
if !gotTight {
|
|
t.Error("tight hit should have been injected")
|
|
}
|
|
}
|
|
|
|
// TestInjectPlaybookMisses_EmptyHits is a fast-path no-op check.
|
|
func TestInjectPlaybookMisses_EmptyHits(t *testing.T) {
|
|
results := []Result{{ID: "w-1", Corpus: "workers", Distance: 0.30}}
|
|
out, injected := InjectPlaybookMisses("test query", results, nil, 0, nil)
|
|
if injected != 0 {
|
|
t.Errorf("expected 0 injection, got %d", injected)
|
|
}
|
|
if len(out) != 1 {
|
|
t.Errorf("results should be unchanged, got len=%d", len(out))
|
|
}
|
|
}
|
|
|
|
func abs(f float64) float64 {
|
|
if f < 0 {
|
|
return -f
|
|
}
|
|
return f
|
|
}
|
|
|
|
func idsOf(rs []Result) []string {
|
|
out := make([]string, len(rs))
|
|
for i, r := range rs {
|
|
out[i] = r.ID
|
|
}
|
|
return out
|
|
}
|