Update consumable pipeline, client API docs, and deployment config

- Refine effective candidate consumables and classifier labels
- Adjust vision algorithm, TSV logging, and video session wiring
- Refresh client surgery HTTP contract doc and staging/video docs
- Update settings, docker-compose prod, tests, and uv.lock

Made-with: Cursor

app/services/video/session_manager.py

@@ -140,7 +140,7 @@ class CameraSessionManager:
         if not resolved:
             raise SurgeryPipelineError(
                 "RECORDING_CANNOT_START",
-                "耗材候选为空：请在请求中传入 candidate_consumables，或配置耗材目录 Excel / 分类模型。",
+                "耗材候选为空：请在请求中传入 candidate_consumables，或提供有效的 consumable_classifier_labels.yaml / 分类模型。",
             )
         if not any(str(x).strip() for x in candidate_consumables):
             logger.info(

app/services/video/session_registry.py

@@ -59,7 +59,7 @@ class CameraStreamInferState:
 @dataclass
 class SurgerySessionState:
     candidate_consumables: list[str]
-    #: 分类类名（归一化） -> 业务物品 id（Excel 产品编码或名称）。
+    #: 分类类名（归一化） -> 业务物品 id（YAML label_id 或类名）
     name_to_code: dict[str, str] = field(default_factory=dict)
     camera_infer: dict[str, CameraStreamInferState] = field(default_factory=dict)
     details: list[SurgeryConsumptionStored] = field(default_factory=list)