mirror of
https://github.com/fawney19/Aether.git
synced 2026-09-02 01:10:23 +08:00
1d9c77522a
- 删除全部 Python 源码 (src/) 及 Alembic 迁移脚本，归档至 _deprecated_py_src/ - 重构 Rust gateway ai_pipeline: 拆分 planner/finalize 模块，新增 contracts/adaptation 层 - 重组 handlers 模块为 admin/public/proxy/internal/shared 子模块结构 - 新增 executor 模块，引入 Rust 原生数据库迁移 (aether-data/migrations) - 简化 CI/Docker 构建流程，移除 base image 二级构建，统一为单一 app image - 移除 Python 相关基础设施文件 (entrypoint.sh, gunicorn_conf.py, Dockerfile.base)
100 lines
3.3 KiB
Python
100 lines
3.3 KiB
Python
from __future__ import annotations
|
|
|
|
from dataclasses import dataclass
|
|
from typing import TYPE_CHECKING, Any
|
|
|
|
from src.services.candidate.policy import FailoverAction
|
|
from src.services.task.execute.exception_classification import CandidateErrorAction
|
|
from src.services.task.request_state import RequestBodyState
|
|
|
|
if TYPE_CHECKING:
|
|
from src.services.task.execute.failure import TaskFailureOperationsService
|
|
|
|
|
|
@dataclass(slots=True)
|
|
class ExecutionErrorTransition:
|
|
"""执行异常后的状态流转决策。"""
|
|
|
|
failover_action: FailoverAction
|
|
max_retries: int | None = None
|
|
|
|
def as_failover_tuple(self) -> tuple[FailoverAction, int | None]:
|
|
return (self.failover_action, self.max_retries)
|
|
|
|
|
|
@dataclass(slots=True)
|
|
class SyncExecutionState:
|
|
"""同步执行阶段状态容器（候选上下文 + 异常上下文）。"""
|
|
|
|
candidate_record_map: dict[tuple[int, int], str]
|
|
request_body_state: RequestBodyState | None
|
|
last_error: Exception | None = None
|
|
last_candidate: Any | None = None
|
|
|
|
def touch_candidate(self, candidate: Any) -> None:
|
|
self.last_candidate = candidate
|
|
|
|
def track_execution_error(self, *, exec_err: Any, candidate: Any) -> None:
|
|
self.last_candidate = candidate
|
|
cause = getattr(exec_err, "cause", None)
|
|
self.last_error = cause if isinstance(cause, Exception) else None
|
|
|
|
def resolve_candidate_record_id(self, *, candidate_index: int, record_id: str | None) -> str:
|
|
if record_id:
|
|
return str(record_id)
|
|
return str(self.candidate_record_map.get((candidate_index, 0), "") or "")
|
|
|
|
def consume_rectify_retry_extension(
|
|
self, *, max_retries_for_candidate: int, retry_index: int
|
|
) -> int | None:
|
|
if not self.request_body_state:
|
|
return None
|
|
if not self.request_body_state.consume_rectified_this_turn():
|
|
return None
|
|
return max(max_retries_for_candidate, retry_index + 2)
|
|
|
|
def raise_classified_error(
|
|
self,
|
|
*,
|
|
fallback_error: Any,
|
|
failure_ops: TaskFailureOperationsService,
|
|
model_name: str,
|
|
api_format: str,
|
|
) -> None:
|
|
if self.last_error is not None:
|
|
failure_ops.attach_metadata_to_error(
|
|
self.last_error,
|
|
self.last_candidate,
|
|
model_name,
|
|
api_format,
|
|
)
|
|
raise self.last_error
|
|
|
|
if isinstance(fallback_error, Exception):
|
|
raise fallback_error
|
|
|
|
raise RuntimeError("execution_error_handler requested raise without exception context")
|
|
|
|
|
|
def resolve_execution_error_transition(
|
|
*,
|
|
action: CandidateErrorAction,
|
|
state: SyncExecutionState,
|
|
max_retries_for_candidate: int,
|
|
retry_index: int,
|
|
) -> ExecutionErrorTransition:
|
|
"""根据异常动作分类，返回 FailoverEngine 可消费的状态流转结果。"""
|
|
if action == CandidateErrorAction.RETRY_CURRENT:
|
|
return ExecutionErrorTransition(
|
|
failover_action=FailoverAction.RETRY,
|
|
max_retries=state.consume_rectify_retry_extension(
|
|
max_retries_for_candidate=max_retries_for_candidate,
|
|
retry_index=retry_index,
|
|
),
|
|
)
|
|
|
|
return ExecutionErrorTransition(
|
|
failover_action=FailoverAction.CONTINUE,
|
|
max_retries=None,
|
|
)
|