Function BenchmarkEncoderCache

modules/text2vec-openai/vectorizer/batch_test.go:135–157

(b *testing.B)

}

func BenchmarkEncoderCache(b *testing.B) {
	client := &fakeBatchClient{}
	logger, _ := test.NewNullLogger()

	v := text2vecbase.New(client,
		batch.NewBatchVectorizer(client, 50*time.Second,
			batch.Settings{MaxObjectsPerBatch: 100, MaxTokensPerBatch: func(cfg moduletools.ClassConfig) int { return 500000 }, MaxTimePerBatch: 10, HasTokenLimit: true, ReturnsRateLimit: true},
			logger, "test"),
		batch.ReturnBatchTokenizer(1, "", false),
	)
	deadline := time.Now().Add(10 * time.Second)
	cfg := &fakeClassConfig{vectorizePropertyName: false, classConfig: map[string]interface{}{"vectorizeClassName": false}}
	b.ResetTimer()
	for i := 0; i < b.N; i++ {
		ctx, cancl := context.WithDeadline(context.Background(), deadline)
		vecs, errs := v.ObjectBatch(
			ctx, []*models.Object{{Class: "Car"}}, []bool{false}, cfg,
		)
		cancl()
		require.Len(b, errs, 0)
		require.Len(b, vecs, 1)
	}
}