refactor: 统一任务框架 Phase 3 - 用 TaskService/FailoverEngine 替代 FallbackOrchestrator

核心重构：
- 移除 FallbackOrchestrator，用 TaskService + FailoverEngine 替代
- TaskService 作为统一入口，支持 SYNC/ASYNC 两种任务模式
- FailoverEngine 实现候选遍历、重试、故障转移逻辑
- 新增 AttemptFunc/AttemptResult 协议，统一尝试结果表示

功能改进：
- 流式响应首字节探测（30s 超时，空流触发故障转移）
- 流式取消归因优化（区分客户端断连 vs 服务端中断）
- 新增 OpenAI Sora 视频取消路由 POST /v1/videos/{task_id}/cancel
- OpenAI 流式请求自动添加 stream_options.include_usage

代码规范：
- 修复 loguru 日志格式（%s → {}）
- 新增 FORMAT_CONVERSION_ENABLED 环境变量说明

测试覆盖：
- test_failover_engine.py: FailoverEngine 单元测试
- test_task_service_async_execute.py: TaskService ASYNC 模式测试
- test_video_cancel_e2e.py: 视频取消端到端测试

src/services/orchestration/__init__.py

@@ -2,7 +2,6 @@
 Orchestration 模块
 
 提供请求编排相关的组件：
-- FallbackOrchestrator: 故障转移编排器，协调请求的完整生命周期
 - CandidateResolver: 候选解析器，负责获取和排序可用的 Provider 组合
 - RequestDispatcher: 请求分发器，负责执行单个候选请求
 - ErrorClassifier: 错误分类器，负责错误分类和处理策略
@@ -10,11 +9,9 @@ Orchestration 模块
 
 from .candidate_resolver import CandidateResolver
 from .error_classifier import ErrorAction, ErrorClassifier
-from .fallback_orchestrator import FallbackOrchestrator
 from .request_dispatcher import RequestDispatcher
 
 __all__ = [
-    "FallbackOrchestrator",
     "CandidateResolver",
     "RequestDispatcher",
     "ErrorClassifier",

src/services/orchestration/candidate_resolver.py

@@ -77,7 +77,7 @@ class CandidateResolver:
         global_model_id: str | None = None
 
         logger.debug(
-            "[CandidateResolver] fetch_candidates starting: model=%s, api_format=%s",
+            "[CandidateResolver] fetch_candidates starting: model={}, api_format={}",
             model_name,
             api_format,
         )
@@ -96,7 +96,7 @@ class CandidateResolver:
             )
 
             logger.debug(
-                "[CandidateResolver] list_all_candidates batch: offset=%d, returned=%d candidates",
+                "[CandidateResolver] list_all_candidates batch: offset={}, returned={} candidates",
                 provider_offset,
                 len(candidates),
             )
@@ -111,7 +111,7 @@ class CandidateResolver:
             provider_offset += provider_batch_size
 
             logger.debug(
-                "[CandidateResolver] fetch_candidates completed: total=%d candidates",
+                "[CandidateResolver] fetch_candidates completed: total={} candidates",
                 len(all_candidates),
             )