feat: add FastAPI HTTP API layer

- Implement FastAPI server with health, collections, ingest, search, and delete endpoints
- Add Pydantic request models for ingest and search
- Add lazy singleton initialization for DB, embedder, and services
- Support environment variable override for data dir and collection name
- Add integration tests with TestClient and temp directory
- Set TRANSFORMERS_OFFLINE/HF_HUB_OFFLINE for offline model loading in tests

tests/test_api.py

@@ -0,0 +1,92 @@
+"""API 端点集成测试."""
+import os
+import tempfile
+
+import pytest
+from fastapi.testclient import TestClient
+
+
+@pytest.fixture
+def client():
+    """创建测试客户端（使用临时目录 + 本地嵌入模型）."""
+    tmpdir = tempfile.mkdtemp()
+    # 设置环境变量使 app 使用测试配置
+    os.environ["MD_VECTOR_DB_DATA_DIR"] = tmpdir
+    os.environ["MD_VECTOR_DB_COLLECTION"] = "test_api"
+    # 离线模式: 模型已缓存, 无需联网
+    os.environ["TRANSFORMERS_OFFLINE"] = "1"
+    os.environ["HF_HUB_OFFLINE"] = "1"
+
+    from src.server.app import app
+
+    with TestClient(app) as c:
+        yield c
+
+
+class TestHealthEndpoint:
+    """健康检查."""
+
+    def test_health_returns_ok(self, client):
+        response = client.get("/api/v1/health")
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+
+
+class TestCollectionEndpoint:
+    """Collection 端点."""
+
+    def test_list_collections(self, client):
+        response = client.get("/api/v1/collections")
+        assert response.status_code == 200
+        data = response.json()
+        assert "collections" in data
+
+
+class TestSearchEndpoint:
+    """搜索端点."""
+
+    def test_search_requires_query(self, client):
+        response = client.post("/api/v1/search", json={})
+        assert response.status_code == 422  # FastAPI 自动校验
+
+    def test_search_empty_collection(self, client):
+        response = client.post(
+            "/api/v1/search",
+            json={"query": "test", "top_k": 5},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert "results" in data
+        assert data["results"] == []
+
+
+class TestIngestEndpoint:
+    """入库端点."""
+
+    def test_ingest_content(self, client):
+        response = client.post(
+            "/api/v1/ingest",
+            json={
+                "content": "# Test\nHello world.",
+                "file_name": "test.md",
+            },
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+        assert data["chunks"] > 0
+
+    def test_ingest_then_search(self, client):
+        """入库后能检索到."""
+        client.post(
+            "/api/v1/ingest",
+            json={"content": "# 配置说明\nChromaDB 配置很简单。", "file_name": "config.md"},
+        )
+        response = client.post(
+            "/api/v1/search",
+            json={"query": "配置", "top_k": 3},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert len(data["results"]) > 0