M.I.M.I.R - Multi-agent Intelligent Memory & Insight Repository

package embed import ( "context" "sync" "sync/atomic" "testing" ) // mockEmbedder tracks calls for testing type mockEmbedder struct { calls int64 batchSize int } func (m *mockEmbedder) Embed(ctx context.Context, text string) ([]float32, error) { atomic.AddInt64(&m.calls, 1) // Return deterministic embedding based on text hash return []float32{float32(len(text)), 0.5, 0.5}, nil } func (m *mockEmbedder) EmbedBatch(ctx context.Context, texts []string) ([][]float32, error) { atomic.AddInt64(&m.calls, int64(len(texts))) m.batchSize = len(texts) results := make([][]float32, len(texts)) for i, text := range texts { results[i] = []float32{float32(len(text)), 0.5, 0.5} } return results, nil } func (m *mockEmbedder) Model() string { return "mock" } func (m *mockEmbedder) Dimensions() int { return 3 } func (m *mockEmbedder) CallCount() int64 { return atomic.LoadInt64(&m.calls) } func TestCachedEmbedder_CacheHit(t *testing.T) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, 100) ctx := context.Background() // First call - cache miss _, err := cached.Embed(ctx, "hello world") if err != nil { t.Fatal(err) } if mock.CallCount() != 1 { t.Errorf("Expected 1 call, got %d", mock.CallCount()) } // Second call with same text - cache hit _, err = cached.Embed(ctx, "hello world") if err != nil { t.Fatal(err) } if mock.CallCount() != 1 { t.Errorf("Expected still 1 call (cache hit), got %d", mock.CallCount()) } // Third call with different text - cache miss _, err = cached.Embed(ctx, "different text") if err != nil { t.Fatal(err) } if mock.CallCount() != 2 { t.Errorf("Expected 2 calls, got %d", mock.CallCount()) } // Verify stats stats := cached.Stats() if stats.Hits != 1 { t.Errorf("Expected 1 hit, got %d", stats.Hits) } if stats.Misses != 2 { t.Errorf("Expected 2 misses, got %d", stats.Misses) } if stats.Size != 2 { t.Errorf("Expected cache size 2, got %d", stats.Size) } } func TestCachedEmbedder_BatchCaching(t *testing.T) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, 100) ctx := context.Background() // Pre-cache one text _, _ = cached.Embed(ctx, "cached") // Batch with mix of cached and new texts := []string{"cached", "new1", "new2"} _, err := cached.EmbedBatch(ctx, texts) if err != nil { t.Fatal(err) } // Should only call base for new texts (2), plus the pre-cache (1) // Total: 1 (pre-cache) + 2 (batch misses) = 3 if mock.CallCount() != 3 { t.Errorf("Expected 3 total calls, got %d", mock.CallCount()) } // Verify batch size was only the misses if mock.batchSize != 2 { t.Errorf("Expected batch of 2 (misses only), got %d", mock.batchSize) } stats := cached.Stats() if stats.Hits != 1 { t.Errorf("Expected 1 hit (from batch), got %d", stats.Hits) } } func TestCachedEmbedder_LRUEviction(t *testing.T) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, 3) // Small cache ctx := context.Background() // Fill cache _, _ = cached.Embed(ctx, "a") _, _ = cached.Embed(ctx, "b") _, _ = cached.Embed(ctx, "c") if cached.Stats().Size != 3 { t.Errorf("Expected size 3, got %d", cached.Stats().Size) } // Add one more - should evict "a" (oldest) _, _ = cached.Embed(ctx, "d") if cached.Stats().Size != 3 { t.Errorf("Expected size still 3 after eviction, got %d", cached.Stats().Size) } // "a" should be evicted, so this should be a cache miss callsBefore := mock.CallCount() _, _ = cached.Embed(ctx, "a") if mock.CallCount() == callsBefore { t.Error("Expected cache miss for evicted 'a', but got hit") } } func TestCachedEmbedder_Concurrent(t *testing.T) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, 1000) ctx := context.Background() var wg sync.WaitGroup for i := 0; i < 100; i++ { wg.Add(1) go func(i int) { defer wg.Done() text := "text" if i%2 == 0 { text = "other" } _, err := cached.Embed(ctx, text) if err != nil { t.Error(err) } }(i) } wg.Wait() // Should only have 2 unique texts cached stats := cached.Stats() if stats.Size != 2 { t.Errorf("Expected 2 unique cached, got %d", stats.Size) } // Should have high hit rate (100 calls, only 2 unique = 98% hits) if stats.HitRate < 90 { t.Errorf("Expected >90%% hit rate, got %.2f%%", stats.HitRate) } } func BenchmarkCachedEmbedder_CacheHit(b *testing.B) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, 1000) ctx := context.Background() // Pre-warm cache _, _ = cached.Embed(ctx, "benchmark text") b.ResetTimer() for i := 0; i < b.N; i++ { _, _ = cached.Embed(ctx, "benchmark text") } } func BenchmarkCachedEmbedder_CacheMiss(b *testing.B) { mock := &mockEmbedder{} cached := NewCachedEmbedder(mock, b.N+1) ctx := context.Background() b.ResetTimer() for i := 0; i < b.N; i++ { text := string(rune('a' + i%26)) _, _ = cached.Embed(ctx, text) } }