Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go

Source identity

Field	Recorded value
Source type	GitHub source
Original record	Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go
Date	Not specified in source
Repository	Janady13/missionmatch
Revision	720d47fae476b0d38477ae4401697d372b1ea2bc
Branch	feature/opportunity-to-outcome
Path or archive member	internal/integration/retrieval_snippet_integration_test.go
Original SHA-256	50c62c8edbc3951c0832cf9244a3533703020f5241c53ad7a53a6dad260351f9
Git blob SHA	d98d969e3e3447a91aabbf555de5bd31b3101f6e
Reading-copy SHA-256	50c62c8edbc3951c0832cf9244a3533703020f5241c53ad7a53a6dad260351f9

Original source ↗

Every recorded occurrence

These 12 records contain identical original source bytes. Their paths and revisions remain part of the archive.

Original record	Branch or archive	Revision / identifier	Original location
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	feature/opportunity-to-outcome	720d47fae476b0d38477ae4401697d372b1ea2bc	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	collision-deconfliction	4ee5a09972608ed89f8377c6665542d5f450f0a6	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	m5-price-to-win	41000ad30fff4129fdaf5ec84e15ba14c27d4890	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	master	1e9249840a4302ec82e123b9c9cb6f75badea3e4	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	mag-substrate	b82c7100a492dec67dfd1eda6b3f53def859be35	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	time-to-fill-survival	b66e1e505863d4f17741a4b9ede73509e45bd6c6	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	graphrag-foundation	b0ddeb9f633fd138d09309d1a17b7d13bc3a67f7	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	presentation-readiness	eb574acaa29d71ce4d0f852572e87c275c79fdb7	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	vendor-intelligence	fe9796a19c261951d3f2565206c0c1b5ac294465	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	phase2-connectors	e038d4c687aec16f8bd9bc4d907eea14d91f8704	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	feature/aioe-upgrade	2178e7298d0e5d9af83afa12402fd587f0ba6ed8	Open original
Janady13/missionmatch / internal/integration/retrieval_snippet_integration_test.go	capture-readiness	bae6af7cb2997e819174a9e9cfff5e16067a1975	Open original

Recovered source transcript

Download complete source text ↓116 lines · 4.8 KB

//go:build integration

// Snippet on the hit, exercised against a REAL Postgres with pgvector, full text, and RLS in force. It
// proves the added ts_headline snippet expression on the hybrid retrieval query: a candidate seeded with a
// distinctive headline term, retrieved under RLS with that term as the text query, comes back with a non
// empty Snippet that contains the matched term, so the faithfulness gate and the narration can cite WHY the
// candidate matched, the matched TEXT, not only the name. It also proves the degrade contract: a pure
// semantic query (an empty text query, so an empty tsquery) still returns hits through the vector arm with
// an empty, no highlight snippet and no error. The snippet is built over the SAME text the search_tsv is
// built from and bound to the SAME tsquery the full text leg already uses, so it never diverges from the
// query that produced the match. This tests the query directly via dbgen with a deterministic embedding, so
// it needs no inference client, exactly like the sibling hybrid search tests.
package integration

import (
	"context"
	"strings"
	"testing"

	"github.com/google/uuid"
	"github.com/jackc/pgx/v5"
	"github.com/pgvector/pgvector-go"

	"github.com/missionworkforce/missionmatch/internal/platform/database/dbgen"
	"github.com/missionworkforce/missionmatch/internal/platform/embeddings"
)

func TestHybridSearchSnippetOnHit(t *testing.T) {
	orgA := seedOrg(t)
	emb := embeddings.NewDeterministic(384)

	// A distinctive headline term that the snippet must surface. It lives in the headline, one of the three
	// fields the search_tsv (and the matching ts_headline source text) is built from in migration 0002.
	const term = "cryptanalysis"
	seed := func(name, headline string) uuid.UUID {
		vec, _ := emb.Embed(context.Background(), name+" "+headline)
		pv := pgvector.NewVector(vec)
		hn := headline
		c, err := dbgen.New(migrator).CreateCandidate(context.Background(), dbgen.CreateCandidateParams{
			OrganizationID: orgA, FullName: name, Headline: &hn, ClearanceLevel: "top_secret",
			ClearancePoly: "none", ClearanceStatus: "active", Available: true, ResumeEmbedding: &pv,
		})
		if err != nil {
			t.Fatalf("seed candidate %q: %v", name, err)
		}
		return c.ID
	}
	wantID := seed("Alice Analyst", "Senior cryptanalysis lead for SIGINT collection")
	seed("Bob Builder", "Kubernetes and AWS platform engineer")

	// A text matching query: the FTS leg matches Alice on the term, so her snippet must surface it.
	var rows []dbgen.SearchCandidatesHybridRow
	qv := pgvector.NewVector(mustEmbed(emb, term))
	err := appDB.RunInTenantTx(ctxOrg(orgA), func(ctx context.Context, tx pgx.Tx) error {
		if _, e := tx.Exec(ctx, "SET LOCAL hnsw.ef_search = 200"); e != nil {
			return e
		}
		var e error
		rows, e = dbgen.New(tx).SearchCandidatesHybrid(ctx, dbgen.SearchCandidatesHybridParams{
			Column1: qv, Limit: 10, WebsearchToTsquery: term,
		})
		return e
	})
	if err != nil {
		t.Fatalf("hybrid search with snippet failed: %v", err)
	}

	var found *dbgen.SearchCandidatesHybridRow
	for i := range rows {
		if rows[i].ID == wantID {
			found = &rows[i]
		}
		if rows[i].OrganizationID != orgA {
			t.Fatalf("hybrid search leaked a foreign tenant row: %s", rows[i].OrganizationID)
		}
	}
	if found == nil {
		t.Fatalf("the term matching candidate %s was not retrieved for query %q", wantID, term)
	}
	if strings.TrimSpace(found.Snippet) == "" {
		t.Fatalf("the text matching hit has an empty snippet; the matched span was not attached")
	}
	if !strings.Contains(strings.ToLower(found.Snippet), term) {
		t.Fatalf("the snippet %q does not contain the matched term %q", found.Snippet, term)
	}

	// The degrade contract: a pure semantic query (an empty text query yields an empty tsquery) must still
	// return hits through the vector arm, with an empty, no highlight snippet, and no error. ts_headline on an
	// empty tsquery returns an unhighlighted head, never an error, so an empty snippet is a valid result.
	var semRows []dbgen.SearchCandidatesHybridRow
	err = appDB.RunInTenantTx(ctxOrg(orgA), func(ctx context.Context, tx pgx.Tx) error {
		if _, e := tx.Exec(ctx, "SET LOCAL hnsw.ef_search = 200"); e != nil {
			return e
		}
		var e error
		semRows, e = dbgen.New(tx).SearchCandidatesHybrid(ctx, dbgen.SearchCandidatesHybridParams{
			Column1: qv, Limit: 10, WebsearchToTsquery: "",
		})
		return e
	})
	if err != nil {
		t.Fatalf("pure semantic (empty query) hybrid search must not error, got: %v", err)
	}
	if len(semRows) == 0 {
		t.Fatal("pure semantic search returned no rows; the vector arm should still recall the seeded candidates")
	}
	for _, r := range semRows {
		if r.OrganizationID != orgA {
			t.Fatalf("semantic search leaked a foreign tenant row: %s", r.OrganizationID)
		}
		if strings.TrimSpace(r.Snippet) != "" {
			t.Fatalf("a pure semantic hit must carry an empty snippet (no tsquery to highlight), got %q for %s", r.Snippet, r.FullName)
		}
	}
}