修改问题

sync_state_machine/common/collection.py

@@ -78,6 +78,14 @@ class DataCollection:
             # node_id 不变但 data_id 变更，清理旧索引
             self._data_id_to_node_id.pop(old_node.data_id, None)
 
+        stale_data_ids = [
+            data_id
+            for data_id, existing_node_id in self._data_id_to_node_id.items()
+            if existing_node_id == node.node_id and data_id != node.data_id
+        ]
+        for stale_data_id in stale_data_ids:
+            self._data_id_to_node_id.pop(stale_data_id, None)
+
         if node.data_id and node.data_id != "":
             existing_node_id = self._data_id_to_node_id.get(node.data_id)
             if existing_node_id is not None and existing_node_id != node.node_id:

sync_state_machine/domain/project_detail/api_handler.py

@@ -315,14 +315,19 @@ class ProjectDetailApiHandler(BaseApiHandler):
 
         existing_by_pair = self._find_existing_by_project_key(project_id=project_id, key=key)
         if existing_by_pair is not None:
-            if loaded_data_id and existing_by_pair.data_id != loaded_data_id:
-                with existing_by_pair.allow_core_state_write("project_detail_pair_rebind"):
-                    existing_by_pair.data_id = loaded_data_id
-            existing_by_pair.set_data(data)
-            existing_by_pair.set_origin_data(data)
-            if project_id:
-                existing_by_pair.context["project_id"] = project_id
-            return existing_by_pair
+            if loaded_data_id and existing_by_pair.data_id == loaded_data_id:
+                existing_by_pair.set_data(data)
+                existing_by_pair.set_origin_data(data)
+                if project_id:
+                    existing_by_pair.context["project_id"] = project_id
+                return existing_by_pair
+
+            existing_data_id = existing_by_pair.data_id or "<empty>"
+            loaded_data_id_text = loaded_data_id or "<empty>"
+            raise RuntimeError(
+                "project_detail_data load conflict: "
+                f"project_id={project_id}, key={key}, existing_data_id={existing_data_id}, loaded_data_id={loaded_data_id_text}"
+            )
 
         node = self._create_basic_node(data, depend_ids=[])
         if project_id:

sync_state_machine/pipeline/full_sync_pipeline.py

@@ -157,10 +157,24 @@ class FullSyncPipeline:
 
                 created = await strategy.create()
                 self._logger.info(f"✅ Strategy create: {node_type} (created={len(created)})")
+                created_local_node_ids = {
+                    node.node_id for node in created if self.local_collection.get_by_node_id(node.node_id) is not None
+                }
+                created_remote_node_ids = {
+                    node.node_id for node in created if self.remote_collection.get_by_node_id(node.node_id) is not None
+                }
                 create_written = await self.commit_creates(node_type)
                 await self.normalize_create_success_to_update_entry(node_type)
                 if create_written:
                     await self._reload_node_type(node_type, reason="create wrote data")
+                    await refresh_bind_data_id(
+                        node_type=node_type,
+                        local_collection=self.local_collection,
+                        remote_collection=self.remote_collection,
+                        binding_manager=self.binding_manager,
+                        local_node_ids=created_local_node_ids,
+                        remote_node_ids=created_remote_node_ids,
+                    )
 
                 updated = await strategy.update()
                 self._logger.info(f"✅ Strategy update: {node_type} (updated={len(updated)})")

sync_state_machine/sync_system/strategy_ops/bind_ops.py

@@ -115,7 +115,9 @@ async def refresh_bind_data_id(
             continue
         remote_node_id = local_to_remote.get(local_node.node_id)
         remote_node = remote_collection.get_by_node_id(remote_node_id) if remote_node_id else None
-        if remote_node and remote_node.data_id:
+        local_status = getattr(local_node.status, "value", None)
+        remote_status = getattr(remote_node.status, "value", None) if remote_node is not None else None
+        if local_status != "FAILED" and remote_node and remote_node.data_id and remote_status != "FAILED":
             local_node.context["bind_data_id"] = remote_node.data_id
         else:
             local_node.context.pop("bind_data_id", None)
@@ -126,7 +128,9 @@ async def refresh_bind_data_id(
             continue
         local_node_id = remote_to_local.get(remote_node.node_id)
         local_node = local_collection.get_by_node_id(local_node_id) if local_node_id else None
-        if local_node and local_node.data_id:
+        remote_status = getattr(remote_node.status, "value", None)
+        local_status = getattr(local_node.status, "value", None) if local_node is not None else None
+        if remote_status != "FAILED" and local_node and local_node.data_id and local_status != "FAILED":
             remote_node.context["bind_data_id"] = local_node.data_id
         else:
             remote_node.context.pop("bind_data_id", None)

tests/unit/project_detail_upsert/test_project_detail_upsert_poll.py

@@ -263,7 +263,7 @@ async def test_create_all_marks_duplicate_project_key_in_same_batch_as_error() -
 
 
 @pytest.mark.asyncio
-async def test_create_node_reuses_existing_node_by_project_key_and_rebinds_data_id() -> None:
+async def test_create_node_raises_on_existing_project_key_with_different_data_id() -> None:
     collection = DataCollection("remote")
 
     handler = ProjectDetailApiHandler()
@@ -278,17 +278,15 @@ async def test_create_node_reuses_existing_node_by_project_key_and_rebinds_data_
     )
     await collection.add(existing)
 
-    created = handler._create_node(
-        {
-            "id": "remote-detail-99",
-            "project_id": "project-1",
-            "key": "plan_construction",
-            "value": [{"date": "2026-01-01", "capacity": 21.0}],
-        }
-    )
-
-    assert created.node_id == "detail-node-old"
-    assert created.data_id == "remote-detail-99"
+    with pytest.raises(RuntimeError, match="project_detail_data load conflict"):
+        handler._create_node(
+            {
+                "id": "remote-detail-99",
+                "project_id": "project-1",
+                "key": "plan_construction",
+                "value": [{"date": "2026-01-01", "capacity": 21.0}],
+            }
+        )
 
 
 @pytest.mark.asyncio

tests/unit/test_collection_project_filter.py

@@ -34,3 +34,20 @@ async def test_collection_filter_by_project_ids() -> None:
     assert collection.get("c1") is not None
     assert collection.get("p2_node") is not None
     assert collection.get("c2") is None
+
+
+@pytest.mark.asyncio
+async def test_collection_update_clears_stale_data_id_index_for_same_node() -> None:
+    register_test_domain()
+
+    collection = DataCollection("local")
+    node = build_project_node("p1_node", "p1", name="P1", code="P1")
+    await collection.add(node)
+
+    with node.allow_core_state_write("test_rebind"):
+        node.data_id = "p2"
+
+    await collection.update(node)
+
+    assert collection.get_by_data_id("test_project", "p1") is None
+    assert collection.get_by_data_id("test_project", "p2") is node

tests/unit/test_datasource_load_continue_on_error.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+import logging
+
 import pytest
 
 from sync_state_machine.common.collection import DataCollection
@@ -95,14 +97,14 @@ class _DS(BaseDataSource):
 
 
 @pytest.mark.asyncio
-async def test_load_all_continue_when_one_handler_fails(capsys) -> None:
+async def test_load_all_continue_when_one_handler_fails(caplog) -> None:
     ds = _DS()
     ds.add_handler(_FailHandler())
     ds.add_handler(_OkHandler())
     ds.set_collection(DataCollection("local"))
 
-    await ds.load_all(order=["fail", "ok"])
+    with caplog.at_level(logging.INFO):
+        await ds.load_all(order=["fail", "ok"])
 
-    out = capsys.readouterr().out
-    assert "[fail] 加载失败" in out
-    assert "Loaded nodes for ok: 0" in out
+    assert "[fail] 加载失败" in caplog.text
+    assert "Loaded nodes for ok: 0" in caplog.text

tests/unit/test_full_sync_pipeline_downstream_reload.py

@@ -49,6 +49,10 @@ class _StubStrategy:
     async def update(self):
         return []
 
+    def normalize_compare_payload(self, data, data_id_map=None):
+        del data_id_map
+        return dict(data or {})
+
 
 class _FakePersistence:
     async def close(self) -> None:

tests/unit/test_post_check_dependency_id_normalization.py

@@ -150,11 +150,11 @@ async def test_post_check_uses_same_id_resolver_as_create_update_for_depend_ids(
 
     result = await evaluate_consistency_for_node_type(
         node_type="construction_task",
-        strategy=strategy,
         local_collection=local_collection,
         remote_collection=remote_collection,
         binding_manager=binding_manager,
         logger=logging.getLogger("sync_state_machine"),
+        normalize_compare_payload=strategy.normalize_compare_payload,
         depend_fields={"contract_id": "contract"},
         compare_to_remote=True,
     )
@@ -237,11 +237,11 @@ async def test_post_check_ignores_only_explicit_configured_fields(tmp_path: Path
 
     result = await evaluate_consistency_for_node_type(
         node_type="contract",
-        strategy=strategy,
         local_collection=local_collection,
         remote_collection=remote_collection,
         binding_manager=binding_manager,
         logger=logging.getLogger("sync_state_machine"),
+        normalize_compare_payload=strategy.normalize_compare_payload,
         compare_to_remote=True,
         ignore_fields=["code"],
     )

tests/unit/test_project_detail_api_handler.py

@@ -94,3 +94,41 @@ async def test_project_detail_create_failure_preserves_push_id_and_biz_id() -> N
     assert result.metadata["biz_id"] == "project-1"
     assert result.metadata["project_id"] == "project-1"
     assert result.metadata["key"] == "production"
+
+
+def test_project_detail_load_conflict_raises_on_same_project_key_different_data_id() -> None:
+    collection = DataCollection("remote")
+
+    project = _ProjectNode(
+        node_id="project-node-1",
+        data_id="project-1",
+        data={"id": "project-1"},
+    )
+    collection._nodes[project.node_id] = project
+
+    existing_detail = ProjectDetailSyncNode(
+        node_id="detail-node-1",
+        data_id="detail-old",
+        data={
+            "id": "detail-old",
+            "project_id": "project-1",
+            "key": "actual_construction",
+            "value": [{"date": "2026-01-01", "capacity": 1.0}],
+        },
+        context={"project_id": "project-1"},
+    )
+    collection._nodes[existing_detail.node_id] = existing_detail
+    collection._data_id_to_node_id[existing_detail.data_id] = existing_detail.node_id
+
+    handler = ProjectDetailApiHandler()
+    handler.set_collection(collection)
+
+    with pytest.raises(RuntimeError, match="project_detail_data load conflict"):
+        handler._create_node(
+            {
+                "id": "detail-new",
+                "project_id": "project-1",
+                "key": "actual_construction",
+                "value": [{"date": "2026-01-02", "capacity": 2.0}],
+            }
+        )