Janady13/missionmatch / internal/integration/ai_phase1_integration_test.go

Source identity

Field	Recorded value
Source type	GitHub source
Original record	Janady13/missionmatch / internal/integration/ai_phase1_integration_test.go
Date	Not specified in source
Repository	Janady13/missionmatch
Revision	720d47fae476b0d38477ae4401697d372b1ea2bc
Branch	feature/opportunity-to-outcome
Path or archive member	internal/integration/ai_phase1_integration_test.go
Original SHA-256	b7e65b8c3e84b1b411ee868ef075237707996d10d8a2d53974334b31b8346e2a
Git blob SHA	18596503cc2a77e3e909b24dae787cd0aabc058c
Reading-copy SHA-256	264f74e9bfb9b72d55ca74b17761d93f42f3dee3c2086385b1c54a31df7907cc

Original source ↗

Recovered source transcript

Download complete source text ↓140 lines · 6.0 KB

Credential-bearing and unrelated personal details were removed from this reading copy. The source identity and original checksum remain recorded.

//go:build integration

package integration

import (
	"context"
	"os"
	"strings"
	"testing"

	"github.com/jackc/pgx/v5"
	"github.com/shopspring/decimal"

	"github.com/missionworkforce/missionmatch/internal/modules/matching"
	"github.com/missionworkforce/missionmatch/internal/modules/matching/engine"
	"github.com/missionworkforce/missionmatch/internal/platform/config"
	"github.com/missionworkforce/missionmatch/internal/platform/database/dbgen"
	"github.com/missionworkforce/missionmatch/internal/platform/embeddings"
)

// TestCrossTenantDatabaseIsolation proves the database per tenant boundary: a connection bound to one
// tenant database physically cannot read another tenant's database. Postgres refuses a cross database
// reference, so isolation does not depend on RLS or on application code being correct. Skips unless two
// tenant DSNs are provided (set MM_TENANT_A_DSN and MM_TENANT_B_DSN to two provisioned databases).
func TestCrossTenantDatabaseIsolation(t *testing.T) {
	aDSN, bDSN := os.Getenv("MM_TENANT_A_DSN"), os.Getenv("MM_TENANT_B_DSN")
	if aDSN == "" || bDSN == "" {
		t.Skip("MM_TENANT_A_DSN / MM_TENANT_B_DSN unset; skipping the database per tenant isolation test")
	}
	ctx := context.Background()
	a, err := pgx.Connect(ctx, aDSN)
	if err != nil {
		t.Fatalf("connect A: %v", err)
	}
	defer a.Close(ctx)

	// The name of database A, so we can attempt a cross database reference from B.
	var aName string
	if err := a.QueryRow(ctx, "SELECT current_database()").Scan(&aName); err != nil {
		t.Fatalf("current_database A: %v", err)
	}
	b, err := pgx.Connect(ctx, bDSN)
	if err != nil {
		t.Fatalf("connect B: %v", err)
	}
	defer b.Close(ctx)
	var bName string
	_ = b.QueryRow(ctx, "SELECT current_database()").Scan(&bName)
	if aName == bName {
		t.Fatalf("A and B resolve to the same database %q; they must be separate tenant databases", aName)
	}

	// From a connection bound to B, reaching into A's tables must fail at the Postgres level.
	_, err = b.Exec(ctx, "SELECT 1 FROM "+aName+".public.candidates LIMIT 1")
	if err == nil || !strings.Contains(strings.ToLower(err.Error()), "cross-database") {
		t.Fatalf("a cross database read from B into A must be refused by Postgres, got: %v", err)
	}

	// No cross database bridge (dblink / postgres_fdw) may be installed in B.
	var bridges int
	if err := b.QueryRow(ctx, "SELECT count(*) FROM pg_extension WHERE extname IN ('dblink','postgres_fdw')").Scan(&bridges); err != nil {
		t.Fatalf("bridge check: %v", err)
	}
	if bridges != 0 {
		t.Fatalf("a tenant database must have no cross database bridge, found %d", bridges)
	}
}

// TestErasureWipesResumeVector proves the right to erasure also wipes the AI vector: after a candidate is
// embedded, anonymize must null resume_embedding so no semantic trace of the person survives.
func TestErasureWipesResumeVector(t *testing.T) {
	org := seedOrg(t)
	cand := seedCandidate(t, org, "secret", "active", true)
	ctx := context.Background()

	// Give the candidate a non null embedding (as a backfill would).
	vec := "[" + strings.TrimSuffix(strings.Repeat("0.05,", 384), ",") + "]"
	if _, err := migrator.Exec(ctx, "UPDATE candidates SET resume_embedding=$1::vector WHERE id=$2", vec, cand.ID); err != nil {
		t.Fatalf("set embedding: %v", err)
	}
	var present bool
	_ = migrator.QueryRow(ctx, "SELECT resume_embedding IS NOT NULL FROM candidates WHERE id=$1", cand.ID).Scan(&present)
	if !present {
		t.Fatal("precondition: the candidate should have an embedding before erasure")
	}

	// Anonymize (the erasure path) and confirm the vector is gone.
	if err := appDB.RunInTenantTx(ctxOrg(org), func(ctx context.Context, tx pgx.Tx) error {
		_, e := dbgen.New(tx).AnonymizeCandidate(ctx, cand.ID)
		return e
	}); err != nil {
		t.Fatalf("anonymize: %v", err)
	}
	var stillThere bool
	_ = migrator.QueryRow(ctx, "SELECT resume_embedding IS NOT NULL FROM candidates WHERE id=$1", cand.ID).Scan(&stillThere)
	if stillThere {
		t.Fatal("ERASURE LEAK: resume_embedding survived anonymize; the person's semantic vector was not wiped")
	}
}

// TestEmbedderUsesRealSemantics proves the skill matcher now uses real semantics through the MEGAMIND
// plane, not the deterministic hash fallback: a true synonym pair scores meaningfully higher with the
// real embedder than with the fake one, so the engine's Skills component is semantic. Skips unless the
// inference plane is configured.
func TestEmbedderUsesRealSemantics(t *testing.T) {
	url := os.Getenv("MM_MEGAMIND_URL")
	if url == "" {
		t.Skip("MM_MEGAMIND_URL unset; skipping the semantic skill match test")
	}
	ctx := context.Background()
	required := engine.RequiredSkill{Name: "Amazon Web Services"}
	cands := []engine.CandidateSkill{{Name: "AWS cloud"}}
	threshold := decimal.RequireFromString("0.50")

	// APIKey IS [credential removed] Without it the provider sends no bearer, the gateway answers 401, and the matcher
	// reports similarity 0 rather than an error, so this test failed claiming the EMBEDDER lacks semantics
	// when the real cause was an unauthenticated call. A swallowed transport error that reads as a bad
	// model score is the most misleading failure shape there is.
	real, err := embeddings.New(config.Embeddings{Provider: config.ProviderMegamind, MegamindURL: url,
		Model: "all-minilm", Dimension: 384, APIKey: os.Getenv("MM_MEGAMIND_TOKEN")})
	if err != nil {
		t.Fatalf("megamind embedder: %v", err)
	}
	fake, _ := embeddings.New(config.Embeddings{Provider: config.ProviderFake, Dimension: 384})

	realMatch, realOK := matching.NewEmbeddingSkillMatcher(real).Match(ctx, required, cands, threshold)
	_, fakeOK := matching.NewEmbeddingSkillMatcher(fake).Match(ctx, required, cands, threshold)

	if !realOK {
		t.Fatalf("the real embedder should semantically match %q to %q above %s, got similarity %s",
			required.Name, cands[0].Name, threshold, realMatch.Similarity)
	}
	if fakeOK {
		t.Fatal("the fake (hash) embedder should NOT semantically match these distinct strings; if it does the test is not proving real semantics")
	}
	if realMatch.Exact {
		t.Fatal("this is a semantic match, not an exact one; Exact must be false")
	}
}