test: 评估系统重构 - 更新测试适配新语义

- 集成测试适配 dataset/run API 路径
- 单元测试补充并发数参数验证
- conftest 新增测试辅助方法

backend/test/integration/api/test_evaluation_router.py

@@ -11,38 +11,38 @@ import pytest
 pytestmark = [pytest.mark.asyncio, pytest.mark.integration]
 
 
-async def _upload_test_benchmark(test_client, admin_headers: dict[str, str], db_id: str) -> tuple[str, str]:
-    benchmark_name = f"pytest_benchmark_{uuid.uuid4().hex[:8]}"
+async def _upload_test_dataset(test_client, admin_headers: dict[str, str], db_id: str) -> tuple[str, str]:
+    dataset_name = f"pytest_dataset_{uuid.uuid4().hex[:8]}"
     line = '{"query":"什么是单元测试？","gold_answer":"用于验证代码行为的自动化测试"}\n'
 
     response = await test_client.post(
-        f"/api/evaluation/databases/{db_id}/benchmarks/upload",
-        data={"name": benchmark_name, "description": "pytest benchmark for download"},
-        files={"file": ("pytest_benchmark.jsonl", line.encode("utf-8"), "application/x-ndjson")},
+        f"/api/evaluation/databases/{db_id}/datasets/upload",
+        data={"name": dataset_name, "description": "pytest dataset for download"},
+        files={"file": ("pytest_dataset.jsonl", line.encode("utf-8"), "application/x-ndjson")},
         headers=admin_headers,
     )
     assert response.status_code == 200, response.text
 
     payload = response.json()
     assert payload.get("message") == "success"
-    benchmark_id = payload.get("data", {}).get("benchmark_id")
-    assert benchmark_id
-    return benchmark_id, line
+    dataset_id = payload.get("data", {}).get("dataset_id")
+    assert dataset_id
+    return dataset_id, line
 
 
-async def test_download_benchmark_requires_admin(test_client, standard_user):
+async def test_download_dataset_requires_admin(test_client, standard_user):
     response = await test_client.get(
-        "/api/evaluation/benchmarks/benchmark_fake/download",
+        "/api/evaluation/datasets/dataset_fake/download",
         headers=standard_user["headers"],
     )
     assert response.status_code == 403
 
 
-async def test_admin_can_download_benchmark(test_client, admin_headers, knowledge_database):
-    benchmark_id, expected_line = await _upload_test_benchmark(test_client, admin_headers, knowledge_database["db_id"])
+async def test_admin_can_download_dataset(test_client, admin_headers, knowledge_database):
+    dataset_id, expected_line = await _upload_test_dataset(test_client, admin_headers, knowledge_database["db_id"])
 
     response = await test_client.get(
-        f"/api/evaluation/benchmarks/{benchmark_id}/download",
+        f"/api/evaluation/datasets/{dataset_id}/download",
         headers=admin_headers,
     )
     assert response.status_code == 200, response.text
@@ -53,9 +53,9 @@ async def test_admin_can_download_benchmark(test_client, admin_headers, knowledg
     assert expected_line.strip() in content
 
 
-async def test_download_benchmark_not_found(test_client, admin_headers):
+async def test_download_dataset_not_found(test_client, admin_headers):
     response = await test_client.get(
-        f"/api/evaluation/benchmarks/benchmark_not_found_{uuid.uuid4().hex[:8]}/download",
+        f"/api/evaluation/datasets/dataset_not_found_{uuid.uuid4().hex[:8]}/download",
         headers=admin_headers,
     )
     assert response.status_code == 404, response.text

backend/test/integration/conftest.py

@@ -34,6 +34,22 @@ HTTP_TIMEOUT = httpx.Timeout(60.0, connect=5.0)
 SANDBOX_CONTAINER_PREFIX = os.getenv("YUXI_SANDBOX_CONTAINER_PREFIX", "yuxi-sandbox")
 
 
+@pytest.fixture(scope="session", autouse=True)
+def ensure_live_api_schema():
+    if not ADMIN_LOGIN or not ADMIN_PASSWORD:
+        return
+
+    async def run_schema_setup() -> None:
+        from yuxi.storage.postgres.manager import pg_manager
+
+        pg_manager.initialize()
+        await pg_manager.create_tables()
+        await pg_manager.ensure_business_schema()
+        await pg_manager.ensure_knowledge_schema()
+
+    anyio.run(run_schema_setup)
+
+
 def _require_admin_credentials() -> tuple[str, str]:
     if not ADMIN_LOGIN or not ADMIN_PASSWORD:
         pytest.skip("Integration credentials are not configured via TEST_USERNAME / TEST_PASSWORD.")

backend/test/unit/knowledge/eval/test_benchmark_generation.py

@@ -1,3 +1,4 @@
+import asyncio
 import os
 from types import SimpleNamespace
 
@@ -11,6 +12,7 @@ from yuxi.knowledge.eval.benchmark_generation import (
     clamp_neighbors_count,
     collect_kb_chunks,
     iter_generated_benchmark_items,
+    normalize_generation_concurrency_count,
     select_neighbor_chunks_by_kb_query,
 )
 
@@ -65,12 +67,40 @@ class NoQueryKnowledgeBase(FakeGenerationKnowledgeBase):
         raise AssertionError("neighbors_count=1 时不应调用 aquery")
 
 
+class TrackingLlm:
+    def __init__(self, content=None, delay=0):
+        self.content = content or '{"query":"问题","gold_answer":"答案","gold_chunk_ids":["anchor_chunk"]}'
+        self.delay = delay
+        self.active_calls = 0
+        self.max_active_calls = 0
+        self.calls = 0
+
+    async def call(self, prompt, stream):
+        self.calls += 1
+        self.active_calls += 1
+        self.max_active_calls = max(self.max_active_calls, self.active_calls)
+        try:
+            if self.delay:
+                await asyncio.sleep(self.delay)
+            return SimpleNamespace(content=self.content)
+        finally:
+            self.active_calls -= 1
+
+
 def test_clamp_neighbors_count():
     assert clamp_neighbors_count(-1) == 0
     assert clamp_neighbors_count(3) == 3
     assert clamp_neighbors_count(11) == 10
 
 
+def test_normalize_generation_concurrency_count():
+    assert normalize_generation_concurrency_count(None) == 10
+    assert normalize_generation_concurrency_count("") == 10
+    assert normalize_generation_concurrency_count(0) == 1
+    assert normalize_generation_concurrency_count(-5) == 1
+    assert normalize_generation_concurrency_count(10000) == 20
+
+
 def test_build_benchmark_generation_prompt_contains_required_schema():
     prompt = build_benchmark_generation_prompt([("chunk_1", "片段内容")])
 
@@ -188,3 +218,65 @@ async def test_iter_generated_benchmark_items_falls_back_to_anchor_when_query_em
 
     assert items == [{"query": "问题", "gold_chunk_ids": ["anchor_chunk"], "gold_answer": "答案"}]
     assert "片段ID=anchor_chunk" in fake_llm.prompts[0]
+
+
+@pytest.mark.asyncio
+async def test_iter_generated_benchmark_items_respects_concurrency_count(monkeypatch):
+    fake_llm = TrackingLlm(delay=0.01)
+    monkeypatch.setattr(benchmark_generation, "select_model", lambda model_spec: fake_llm)
+
+    items = [
+        item
+        async for item in iter_generated_benchmark_items(
+            kb_instance=NoQueryKnowledgeBase(),
+            db_id="db_1",
+            count=4,
+            neighbors_count=1,
+            concurrency_count=2,
+            llm_model_spec="test-provider:test-model",
+        )
+    ]
+
+    assert len(items) == 4
+    assert fake_llm.max_active_calls == 2
+
+
+@pytest.mark.asyncio
+async def test_iter_generated_benchmark_items_returns_at_most_count(monkeypatch):
+    fake_llm = TrackingLlm(delay=0.01)
+    monkeypatch.setattr(benchmark_generation, "select_model", lambda model_spec: fake_llm)
+
+    items = [
+        item
+        async for item in iter_generated_benchmark_items(
+            kb_instance=NoQueryKnowledgeBase(),
+            db_id="db_1",
+            count=3,
+            neighbors_count=1,
+            concurrency_count=10,
+            llm_model_spec="test-provider:test-model",
+        )
+    ]
+
+    assert len(items) == 3
+
+
+@pytest.mark.asyncio
+async def test_iter_generated_benchmark_items_stops_at_max_attempts(monkeypatch):
+    fake_llm = TrackingLlm(content='{"query":"","gold_answer":"答案","gold_chunk_ids":["anchor_chunk"]}')
+    monkeypatch.setattr(benchmark_generation, "select_model", lambda model_spec: fake_llm)
+
+    items = [
+        item
+        async for item in iter_generated_benchmark_items(
+            kb_instance=NoQueryKnowledgeBase(),
+            db_id="db_1",
+            count=2,
+            neighbors_count=1,
+            concurrency_count=10,
+            llm_model_spec="test-provider:test-model",
+        )
+    ]
+
+    assert items == []
+    assert fake_llm.calls == 50