67 lines
1.6 KiB
Go
67 lines
1.6 KiB
Go
package ctxmgr_test
|
|
|
|
import (
|
|
"context"
|
|
"math"
|
|
"testing"
|
|
|
|
"chat-agent-system/internal/ctxmgr"
|
|
)
|
|
|
|
func TestDeterministicEmbedder_EmbeddingProperties(t *testing.T) {
|
|
emb := ctxmgr.NewDeterministicEmbedder(1024)
|
|
ctx := context.Background()
|
|
|
|
texts := []string{
|
|
"超声软组织切割止血手术设备",
|
|
"微创外科硬管内窥镜摄像系统",
|
|
"超声软组织切割止血手术设备", // 重复文本验证确定性
|
|
}
|
|
|
|
vecs, err := emb.Embed(ctx, texts)
|
|
if err != nil {
|
|
t.Fatalf("embed failed: %v", err)
|
|
}
|
|
if len(vecs) != 3 {
|
|
t.Fatalf("expected 3 vectors, got %d", len(vecs))
|
|
}
|
|
|
|
// 1. 验证维度
|
|
for i, v := range vecs {
|
|
if len(v) != 1024 {
|
|
t.Fatalf("vector %d length expected 1024, got %d", i, len(v))
|
|
}
|
|
// 验证单位化模长 (L2 norm ≈ 1.0)
|
|
var sumSq float64
|
|
for _, val := range v {
|
|
sumSq += float64(val * val)
|
|
}
|
|
norm := math.Sqrt(sumSq)
|
|
if math.Abs(norm-1.0) > 1e-4 {
|
|
t.Fatalf("vector %d not normalized, norm=%f", i, norm)
|
|
}
|
|
}
|
|
|
|
// 2. 验证确定性：相同文本向量完全一致
|
|
for i := 0; i < 1024; i++ {
|
|
if vecs[0][i] != vecs[2][i] {
|
|
t.Fatalf("deterministic mismatch at index %d: %v vs %v", i, vecs[0][i], vecs[2][i])
|
|
}
|
|
}
|
|
|
|
// 3. 验证语义区别：不同文本余弦相似度应小于 0.95
|
|
var dotProduct float32
|
|
for i := 0; i < 1024; i++ {
|
|
dotProduct += vecs[0][i] * vecs[1][i]
|
|
}
|
|
if dotProduct > 0.95 {
|
|
t.Fatalf("distinct texts should not have near-identical embeddings: dotProduct=%f", dotProduct)
|
|
}
|
|
|
|
// 4. 验证 VectorLiteral
|
|
lit := ctxmgr.VectorLiteral(vecs[0][:3])
|
|
if lit == "" || lit[0] != '[' || lit[len(lit)-1] != ']' {
|
|
t.Fatalf("invalid vector literal: %s", lit)
|
|
}
|
|
}
|