Function TestBulkLoadVectorIndex

systest/vector/load_test.go:89–204

(t *testing.T)

}

func TestBulkLoadVectorIndex(t *testing.T) {
	// Step 1: Create a source cluster and load vectors into it
	sourceConf := dgraphtest.NewClusterConfig().
		WithNumAlphas(1).
		WithNumZeros(1).
		WithReplicas(1).
		WithACL(time.Hour)
	sourceCluster, err := dgraphtest.NewLocalCluster(sourceConf)
	require.NoError(t, err)
	defer func() { sourceCluster.Cleanup(t.Failed()) }()
	require.NoError(t, sourceCluster.Start())

	gc, cleanup, err := sourceCluster.Client()
	require.NoError(t, err)
	defer cleanup()
	require.NoError(t, gc.LoginIntoNamespace(context.Background(),
		dgraphapi.DefaultUser, dgraphapi.DefaultPassword, x.RootNamespace))

	hc, err := sourceCluster.HTTPClient()
	require.NoError(t, err)
	require.NoError(t, hc.LoginIntoNamespace(dgraphapi.DefaultUser,
		dgraphapi.DefaultPassword, x.RootNamespace))

	// Set up vector schema and load vectors
	require.NoError(t, gc.SetupSchema(testSchema))

	numVectors := 1000
	pred := "project_description_v"
	rdfs, vectors := dgraphapi.GenerateRandomVectors(0, numVectors, 10, pred)

	mu := &api.Mutation{SetNquads: []byte(rdfs), CommitNow: true}
	_, err = gc.Mutate(mu)
	require.NoError(t, err)

	// Verify vectors are loaded and queryable in source cluster
	for _, vector := range vectors[:5] { // Test first 5 vectors
		similarVectors, err := gc.QueryMultipleVectorsUsingSimilarTo(vector, pred, 5)
		require.NoError(t, err)
		require.GreaterOrEqual(t, len(similarVectors), 3, "similar_to query should return results")
	}

	// Step 2: Export the data from source cluster
	require.NoError(t, hc.Export(dgraphtest.DefaultExportDir, "rdf", -1))

	// Step 3: Set up a cluster for bulk loading and run bulk load on exported data
	bulkOutDir := t.TempDir()
	bulkConf := dgraphtest.NewClusterConfig().
		WithNumAlphas(1).
		WithNumZeros(1).
		WithReplicas(1).
		WithACL(time.Hour).
		WithBulkLoadOutDir(bulkOutDir)

	bulkCluster, err := dgraphtest.NewLocalCluster(bulkConf)
	require.NoError(t, err)
	defer func() { bulkCluster.Cleanup(t.Failed()) }()

	// Start only Zero for bulk loading