feat(server): add workbench API, DSH planner worker and RAG pipeline

server/tests/test_quality_and_rag.py

@@ -0,0 +1,69 @@
+import unittest
+from pathlib import Path
+from tempfile import TemporaryDirectory
+from sqlmodel import SQLModel, Session, create_engine
+
+from app.agents.composer import validate_consistency_checks
+from app.agents.coordinator import provider_fingerprint
+from app.agents.qc import validate_qc_checks
+from app.models import ModelProvider
+from app.rag.pipeline import _disk_files
+from app.rag.store import _as_list, _chunk_id
+from app.rag.pipeline import release_reserved_chunks, reserve_chunks
+from array import array
+
+
+class QualityAndRagTests(unittest.TestCase):
+    def test_empty_qc_is_invalid(self):
+        checks, error = validate_qc_checks([])
+        self.assertEqual(checks, [])
+        self.assertTrue(error)
+
+    def test_qc_requires_named_status(self):
+        checks, error = validate_qc_checks([{"id": "x", "issues": []}])
+        self.assertEqual(checks, [])
+        self.assertTrue(error)
+
+    def test_composer_requires_all_five_fixed_checks(self):
+        checks, error = validate_consistency_checks([])
+        self.assertEqual(checks, [])
+        self.assertTrue(error)
+
+    def test_chunk_id_is_stable(self):
+        item = {"text": "  同一段素材  ", "meta": {"heading": "心法"}}
+        self.assertEqual(_chunk_id("素材.md", item, 1), _chunk_id("素材.md", item, 1))
+        self.assertNotEqual(_chunk_id("素材.md", item, 1), _chunk_id("素材.md", item, 2))
+
+    def test_missing_source_directory_is_an_empty_inventory(self):
+        with TemporaryDirectory() as temp:
+            missing = Path(temp) / "not-created"
+            self.assertEqual(_disk_files(missing), [])
+
+    def test_provider_fingerprint_changes_when_endpoint_changes(self):
+        provider = ModelProvider(id=7, name="gateway", base_url="https://a.example/v1")
+        first = provider_fingerprint(provider, "model-a")
+        self.assertEqual(first, provider_fingerprint(provider, "model-a"))
+        provider.base_url = "https://b.example/v1"
+        self.assertNotEqual(first, provider_fingerprint(provider, "model-a"))
+
+    def test_vector_snapshot_sequences_are_numpy_safe(self):
+        values = array("f", [0.1, 0.2])
+        self.assertEqual(len(_as_list(values)), 2)
+        self.assertAlmostEqual(_as_list(values)[0], 0.1, places=6)
+        self.assertAlmostEqual(_as_list(values)[1], 0.2, places=6)
+        self.assertEqual(_as_list(None), [])
+
+    def test_knowledge_usage_is_unique_per_agent_and_chunk(self):
+        engine = create_engine("sqlite://")
+        SQLModel.metadata.create_all(engine)
+        with Session(engine) as session:
+            hit = {"id": "chunk-1", "kb_id": 3, "text": "素材", "meta": {"file": "a.md"}}
+            self.assertEqual(len(reserve_chunks(session, 10, 20, 30, [hit])), 1)
+            self.assertEqual(len(reserve_chunks(session, 10, 20, 31, [hit])), 0)
+            self.assertEqual(len(reserve_chunks(session, 11, 20, 31, [hit])), 1)
+            release_reserved_chunks(session, 10, 20, 30, [hit])
+            self.assertEqual(len(reserve_chunks(session, 10, 20, 32, [hit])), 1)
+
+
+if __name__ == "__main__":
+    unittest.main()