File size: 5,814 Bytes
8baf122 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 | package semanticcache
import (
"testing"
"time"
"github.com/maximhq/bifrost/core/schemas"
)
// TestEmbeddingRequestsCaching tests that embedding requests are properly cached using direct hash matching
func TestEmbeddingRequestsCaching(t *testing.T) {
setup := NewTestSetup(t)
defer setup.Cleanup()
ctx := CreateContextWithCacheKey("test-embedding-cache")
// Create embedding request
embeddingRequest := CreateEmbeddingRequest([]string{
"What is machine learning?",
"Explain artificial intelligence in simple terms.",
})
t.Log("Making first embedding request (should go to OpenAI and be cached)...")
// Make first request (will go to OpenAI and be cached) - with retries
start1 := time.Now()
response1, err1 := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
duration1 := time.Since(start1)
if err1 != nil {
return // Test will be skipped by retry function
}
if response1 == nil || len(response1.Data) == 0 {
t.Fatal("First embedding response is invalid")
}
t.Logf("First embedding request completed in %v", duration1)
t.Logf("Response contains %d embeddings", len(response1.Data))
// Wait for cache to be written
WaitForCache(setup.Plugin)
t.Log("Making second identical embedding request (should be served from cache)...")
// Make second identical request (should be cached)
start2 := time.Now()
response2, err2 := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
duration2 := time.Since(start2)
if err2 != nil {
t.Fatalf("Second embedding request failed: %v", err2)
}
if response2 == nil || len(response2.Data) == 0 {
t.Fatal("Second embedding response is invalid")
}
// Verify cache hit
AssertCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response2}, "direct")
t.Logf("Second embedding request completed in %v", duration2)
// Cache should be significantly faster
if duration2 >= duration1 { // Allow some margin but cache should be much faster
t.Log("⚠️ Cache doesn't seem faster, but this could be due to test environment")
}
// Responses should be identical
if len(response1.Data) != len(response2.Data) {
t.Errorf("Response lengths differ: %d vs %d", len(response1.Data), len(response2.Data))
}
t.Log("✅ Embedding requests properly cached using direct hash matching")
}
// TestEmbeddingRequestsNoCacheWithoutCacheKey tests that embedding requests without cache key are not cached
func TestEmbeddingRequestsNoCacheWithoutCacheKey(t *testing.T) {
setup := NewTestSetup(t)
defer setup.Cleanup()
// Don't set cache key in context
ctx := CreateContextWithCacheKey("")
embeddingRequest := CreateEmbeddingRequest([]string{"Test embedding without cache key"})
t.Log("Making embedding request without cache key...")
response, err := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
if err != nil {
t.Fatalf("Embedding request failed: %v", err)
}
// Should not be cached
AssertNoCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response})
t.Log("✅ Embedding requests without cache key are properly not cached")
}
// TestEmbeddingRequestsDifferentTexts tests that different embedding texts produce different cache entries
func TestEmbeddingRequestsDifferentTexts(t *testing.T) {
setup := NewTestSetup(t)
defer setup.Cleanup()
ctx := CreateContextWithCacheKey("test-embedding-different")
// Create two different embedding requests
request1 := CreateEmbeddingRequest([]string{"First set of texts"})
request2 := CreateEmbeddingRequest([]string{"Second set of texts"})
t.Log("Making first embedding request...")
response1, err1 := setup.Client.EmbeddingRequest(ctx, request1)
if err1 != nil {
return // Test will be skipped by retry function
}
AssertNoCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response1})
WaitForCache(setup.Plugin)
t.Log("Making second different embedding request...")
response2, err2 := setup.Client.EmbeddingRequest(ctx, request2)
if err2 != nil {
return // Test will be skipped by retry function
}
// Should not be a cache hit since texts are different
AssertNoCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response2})
t.Log("✅ Different embedding texts produce different cache entries")
}
// TestEmbeddingRequestsCacheExpiration tests TTL functionality for embedding requests
func TestEmbeddingRequestsCacheExpiration(t *testing.T) {
setup := NewTestSetup(t)
defer setup.Cleanup()
// Set very short TTL for testing
shortTTL := 5 * time.Second
ctx := CreateContextWithCacheKeyAndTTL("test-embedding-ttl", shortTTL)
embeddingRequest := CreateEmbeddingRequest([]string{"TTL test embedding"})
t.Log("Making first embedding request with short TTL...")
response1, err1 := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
if err1 != nil {
return // Test will be skipped by retry function
}
AssertNoCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response1})
WaitForCache(setup.Plugin)
t.Log("Making second request before TTL expiration...")
response2, err2 := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
if err2 != nil {
if err2.Error != nil {
t.Fatalf("Second request failed: %v", err2.Error.Message)
} else {
t.Fatalf("Second request failed: %v", err2)
}
}
AssertCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response2}, "direct")
t.Logf("Waiting for TTL expiration (%v)...", shortTTL)
time.Sleep(shortTTL + 1*time.Second) // Wait for TTL to expire
t.Log("Making third request after TTL expiration...")
response3, err3 := setup.Client.EmbeddingRequest(ctx, embeddingRequest)
if err3 != nil {
return // Test will be skipped by retry function
}
// Should not be a cache hit since TTL expired
AssertNoCacheHit(t, &schemas.BifrostResponse{EmbeddingResponse: response3})
t.Log("✅ Embedding requests properly handle TTL expiration")
}
|