73 lines
2.5 KiB
Python
73 lines
2.5 KiB
Python
"""嵌入模型测试."""
|
|
import pytest
|
|
|
|
from src.core.config import EmbedConfig
|
|
from src.core.embedder import LocalEmbedder, APIEmbedder, create_embedder, batch_embed
|
|
|
|
|
|
class TestLocalEmbedder:
|
|
"""本地嵌入器测试."""
|
|
|
|
@pytest.fixture
|
|
def local_config(self):
|
|
return EmbedConfig(mode="local")
|
|
|
|
def test_create_local_embedder(self, local_config):
|
|
"""创建本地嵌入器，验证维度正确."""
|
|
embedder = create_embedder(local_config)
|
|
assert embedder.dimension > 0
|
|
assert isinstance(embedder.dimension, int)
|
|
assert isinstance(embedder, LocalEmbedder)
|
|
|
|
def test_embed_single_text(self, local_config):
|
|
"""嵌入单条文本返回正确维度向量."""
|
|
embedder = create_embedder(local_config)
|
|
result = embedder.embed(["你好世界"])
|
|
assert len(result) == 1
|
|
assert len(result[0]) == embedder.dimension
|
|
assert all(isinstance(v, float) for v in result[0])
|
|
|
|
def test_embed_multiple_texts(self, local_config):
|
|
"""嵌入多条文本返回对应数量的向量."""
|
|
embedder = create_embedder(local_config)
|
|
texts = ["第一段文本", "第二段文本", "第三段文本"]
|
|
result = embedder.embed(texts)
|
|
assert len(result) == 3
|
|
for vec in result:
|
|
assert len(vec) == embedder.dimension
|
|
|
|
def test_embed_empty_list_raises(self, local_config):
|
|
"""空列表应抛出异常."""
|
|
embedder = create_embedder(local_config)
|
|
with pytest.raises(ValueError):
|
|
embedder.embed([])
|
|
|
|
|
|
class TestAPIEmbedder:
|
|
"""API 嵌入器测试."""
|
|
|
|
def test_api_embedder_init(self):
|
|
"""API 嵌入器初始化."""
|
|
cfg = EmbedConfig(mode="api", api_base="https://api.test.com", api_key="sk-test")
|
|
emb = APIEmbedder(cfg)
|
|
assert emb.dimension == 1536
|
|
|
|
def test_api_embedder_empty_raises(self):
|
|
"""API 嵌入器空列表抛异常."""
|
|
cfg = EmbedConfig(mode="api")
|
|
emb = APIEmbedder(cfg)
|
|
with pytest.raises(ValueError):
|
|
emb.embed([])
|
|
|
|
|
|
class TestBatchEmbed:
|
|
"""分批嵌入测试."""
|
|
|
|
def test_batch_embed(self):
|
|
"""分批嵌入返回正确数量."""
|
|
embedder = create_embedder(EmbedConfig(mode="local"))
|
|
texts = ["测试文本"] * 70 # 超过 batch_size 32
|
|
results = batch_embed(embedder, texts, batch_size=32)
|
|
assert len(results) == 70
|
|
assert len(results[0]) == embedder.dimension
|