Behavision/server/internal/store/api_faces_live_test.go

package store

import (
	"context"
	"fmt"
	"testing"
	"time"

	"github.com/loyaly/behavision-server/internal/contract"
	"github.com/loyaly/behavision-server/internal/ingest"
)

// Face images held by this server, against a real database.
//
// The prune is the whole reason this is allowed to live in Postgres at all -
// migration 011 argues it explicitly against 009's "face images grow with every
// visitor who ever walks in" - so it is the one behaviour that must be proved
// against the real thing rather than a fake that would simply agree with me.

func seedAgentSite(t *testing.T, st *Store, name string) ingest.Site {
	t.Helper()
	ctx := context.Background()
	var site ingest.Site
	if err := st.pool.QueryRow(ctx, `
		INSERT INTO clients (name, slug) VALUES ($1, $1) RETURNING id::text`,
		name).Scan(&site.ClientID); err != nil {
		t.Fatalf("seed client: %v", err)
	}
	dropTenant(t, st, site.ClientID)
	if err := st.pool.QueryRow(ctx, `
		INSERT INTO sites (client_id, name, slug) VALUES ($1::uuid, $2, $3)
		RETURNING id::text`, site.ClientID, name, name).Scan(&site.SiteID); err != nil {
		t.Fatalf("seed site: %v", err)
	}
	if err := st.pool.QueryRow(ctx, `
		INSERT INTO agents (client_id, site_id, mqtt_username)
		VALUES ($1::uuid, $2::uuid, $3)
		RETURNING id::text`, site.ClientID, site.SiteID, name).Scan(&site.AgentID); err != nil {
		t.Fatalf("seed agent: %v", err)
	}
	site.Slug = name
	return site
}

func embedding(seed float32) []float32 {
	v := make([]float32, contract.EmbeddingDim)
	for i := range v {
		v[i] = seed
	}
	return v
}

// The bound: one person seen many times leaves ONE stored image, not one per
// visit. Without this the table grows with footfall, which is precisely the
// property that keeps face images out of the database everywhere else.
func TestLiveOnlyOneFaceSurvivesPerVisitor(t *testing.T) {
	st := liveStore(t)
	ctx := context.Background()
	site := seedAgentSite(t, st, "faces-"+stamp())

	var keys []string
	for i := 0; i < 5; i++ {
		key, err := st.PutVisitFace(ctx, site.ClientID, site.SiteID,
			[]byte(fmt.Sprintf("jpeg-%d", i)))
		if err != nil {
			t.Fatalf("store face %d: %v", i, err)
		}
		keys = append(keys, key)

		// The SAME person every time: one embedding, so the matcher resolves
		// them to one visitor.
		ok, err := st.RecordVisit(ctx, site, &contract.Visit{
			EventID:    fmt.Sprintf("%s-%d", site.Slug, i),
			OccurredAt: time.Now().UTC().Add(time.Duration(i) * time.Second),
			CameraID:   "door",
			IsNew:      i == 0,
			Quality:    0.8,
			Similarity: 0.9,
			Embedding:  embedding(0.05),
			ImageKey:   key,
		})
		if err != nil || !ok {
			t.Fatalf("visit %d: ok=%v err=%v", i, ok, err)
		}
	}

	var stored int
	if err := st.pool.QueryRow(ctx,
		`SELECT count(*) FROM visit_faces WHERE client_id = $1::uuid`,
		site.ClientID).Scan(&stored); err != nil {
		t.Fatal(err)
	}
	if stored != 1 {
		t.Fatalf("five visits by one person left %d stored faces - the table "+
			"grows with footfall, which is exactly what migration 011 promises "+
			"it does not", stored)
	}

	// And it is the NEWEST that survived: every surface shows a customer's
	// latest view, so keeping an older one would quietly show a stale face.
	var surviving string
	if err := st.pool.QueryRow(ctx,
		`SELECT 'db:' || id::text FROM visit_faces WHERE client_id = $1::uuid`,
		site.ClientID).Scan(&surviving); err != nil {
		t.Fatal(err)
	}
	if surviving != keys[len(keys)-1] {
		t.Errorf("kept %s, want the newest %s", surviving, keys[len(keys)-1])
	}

	// The superseded keys are blanked, not left dangling. A visit advertising
	// an image that is not there renders as a broken picture on the one screen
	// meant to show it.
	var dangling int
	if err := st.pool.QueryRow(ctx, `
		SELECT count(*) FROM visits v
		 WHERE v.client_id = $1::uuid AND v.image_key LIKE 'db:%'
		   AND NOT EXISTS (SELECT 1 FROM visit_faces f
		                    WHERE 'db:' || f.id::text = v.image_key)`,
		site.ClientID).Scan(&dangling); err != nil {
		t.Fatal(err)
	}
	if dangling != 0 {
		t.Errorf("%d visits point at a face that is gone", dangling)
	}

	// image_deleted_at is the record of an ERASURE and is what an auditor
	// reads. Ordinary housekeeping must not write into it.
	var marked int
	if err := st.pool.QueryRow(ctx, `
		SELECT count(*) FROM visits
		 WHERE client_id = $1::uuid AND image_deleted_at IS NOT NULL`,
		site.ClientID).Scan(&marked); err != nil {
		t.Fatal(err)
	}
	if marked != 0 {
		t.Errorf("%d visits were marked as erased by a routine prune", marked)
	}
}

// Two different people keep one face each. The prune must be scoped to the
// person, not to the site - otherwise every new arrival would delete the
// previous customer's photo.
func TestLiveThePruneIsPerPersonNotPerSite(t *testing.T) {
	st := liveStore(t)
	ctx := context.Background()
	site := seedAgentSite(t, st, "faces2-"+stamp())

	for i, seed := range []float32{0.05, -0.05} {
		key, err := st.PutVisitFace(ctx, site.ClientID, site.SiteID,
			[]byte(fmt.Sprintf("person-%d", i)))
		if err != nil {
			t.Fatal(err)
		}
		if ok, err := st.RecordVisit(ctx, site, &contract.Visit{
			EventID:    fmt.Sprintf("%s-p%d", site.Slug, i),
			OccurredAt: time.Now().UTC(),
			CameraID:   "door",
			IsNew:      true,
			Quality:    0.8,
			Embedding:  embedding(seed),
			ImageKey:   key,
		}); err != nil || !ok {
			t.Fatalf("visit: ok=%v err=%v", ok, err)
		}
	}

	var stored int
	if err := st.pool.QueryRow(ctx,
		`SELECT count(*) FROM visit_faces WHERE client_id = $1::uuid`,
		site.ClientID).Scan(&stored); err != nil {
		t.Fatal(err)
	}
	if stored != 2 {
		t.Fatalf("two people should keep one face each, got %d", stored)
	}
}

// An agent uploads a face BEFORE the server has decided who it is, so a row is
// briefly unreferenced by design - and permanently so if the visit that would
// have claimed it never arrives. That is a stored photograph of a real person
// that nothing points at, which erasure could never reach because it is found
// through the visitor and this row has none.
func TestLiveAnUnclaimedFaceIsSweptAway(t *testing.T) {
	st := liveStore(t)
	ctx := context.Background()
	site := seedAgentSite(t, st, "faces3-"+stamp())

	key, err := st.PutVisitFace(ctx, site.ClientID, site.SiteID, []byte("orphan"))
	if err != nil {
		t.Fatal(err)
	}
	// Age it past the sweep window rather than sleeping.
	if _, err := st.pool.Exec(ctx, `
		UPDATE visit_faces SET captured_at = now() - interval '3 days'
		 WHERE 'db:' || id::text = $1`, key); err != nil {
		t.Fatal(err)
	}

	n, err := st.SweepOrphanFaces(ctx, "1 day")
	if err != nil {
		t.Fatalf("sweep: %v", err)
	}
	if n < 1 {
		t.Fatal("the orphan was not swept")
	}
	if _, err := st.VisitFace(ctx, site.ClientID, key); err == nil {
		t.Fatal("the orphan is still readable")
	}
}

// A face a visit DOES point at must survive the sweep, however old it is. A
// regular customer's photo is exactly the row that gets old.
func TestLiveTheSweepKeepsAClaimedFace(t *testing.T) {
	st := liveStore(t)
	ctx := context.Background()
	site := seedAgentSite(t, st, "faces4-"+stamp())

	key, err := st.PutVisitFace(ctx, site.ClientID, site.SiteID, []byte("kept"))
	if err != nil {
		t.Fatal(err)
	}
	if ok, err := st.RecordVisit(ctx, site, &contract.Visit{
		EventID: site.Slug + "-keep", OccurredAt: time.Now().UTC(),
		CameraID: "door", IsNew: true, Quality: 0.8,
		Embedding: embedding(0.07), ImageKey: key,
	}); err != nil || !ok {
		t.Fatalf("visit: ok=%v err=%v", ok, err)
	}
	if _, err := st.pool.Exec(ctx, `
		UPDATE visit_faces SET captured_at = now() - interval '400 days'
		 WHERE 'db:' || id::text = $1`, key); err != nil {
		t.Fatal(err)
	}

	if _, err := st.SweepOrphanFaces(ctx, "1 day"); err != nil {
		t.Fatal(err)
	}
	if _, err := st.VisitFace(ctx, site.ClientID, key); err != nil {
		t.Fatalf("a claimed face was swept away: %v", err)
	}
}

// An image key travels in API responses. A caller who kept one, or guessed one,
// must get nothing rather than another company's customer.
func TestLiveAFaceIsNotReadableByAnotherTenant(t *testing.T) {
	st := liveStore(t)
	ctx := context.Background()
	a := seedAgentSite(t, st, "facesa-"+stamp())
	b := seedAgentSite(t, st, "facesb-"+stamp())

	key, err := st.PutVisitFace(ctx, a.ClientID, a.SiteID, []byte("private"))
	if err != nil {
		t.Fatal(err)
	}
	if _, err := st.VisitFace(ctx, b.ClientID, key); err == nil {
		t.Fatal("another tenant read a stored face")
	}
	if _, err := st.VisitFace(ctx, a.ClientID, key); err != nil {
		t.Fatalf("the owning tenant could not read its own face: %v", err)
	}
}