Files
xiaoxia-saas/packages/domain/generation_task.py
T
用户CI Test f5f54b98b6
CI/CD Pipeline / Validate Code Quality And Tests (pull_request) Failing after 8s
CI/CD Pipeline / Production Browser E2E (pull_request) Failing after 1655h53m46s
CI/CD Pipeline / Staging E2E Tests (pull_request) Failing after 1655h53m48s
CI/CD Pipeline / Build Production Runtime Images (pull_request) Failing after 1655h53m50s
CI/CD Pipeline / Deploy Production (pull_request) Failing after 1655h53m48s
CI/CD Pipeline / Build & Push Staging (Watchtower auto-deploy) (pull_request) Failing after 1655h53m50s
CI/CD Pipeline / Frontend Lint (pull_request) Has been skipped
CI/CD Pipeline / Staging API Integration Tests (pull_request) Failing after 1656h25m18s
fix(generation): 修复生成任务状态机缺失导致状态永不更新的问题
## 问题根因
generate_video Celery 任务完全没有状态更新逻辑:
- 任务开始执行时,不更新 GenerationTask 状态为 running
- 任务成功时,不更新状态为 completed
- 任务失败时,不更新状态为 failed
- GenerationTask 领域模型缺少 mark_processing/mark_completed/mark_failed 方法

结果:任务状态永远停留在 pending,前端查询时一直显示"生成中"或超时显示失败,即使视频实际上已经生成成功了。

## 修复方案

### 1. GenerationTask 领域模型增加完整状态机
- 新增 TERMINAL_STATUSES 终态集合
- 新增 _VALID_TRANSITIONS 合法状态转换表
- 新增 transition_to() 通用状态转换方法(含合法性校验)
- 新增 mark_processing() — pending → running,设置 started_at
- 新增 mark_completed(result_count) — running → completed,设置 completed_at/progress/result_count
- 新增 mark_failed(error_message) — pending/running → failed,设置 error_message/completed_at
- 新增 mark_cancelled() — pending/running → cancelled
- 新增 mark_pending_from_failed() — failed → pending(用于重试)
- 新增 is_terminal/is_completed/is_failed/is_running 属性

状态机:
  pending → running → completed
              ↘ failed → pending (重试)
       ↘ cancelled

### 2. generate_video Celery 任务增加状态更新
- 新增 _update_task_status() 辅助函数:独立 session + 异常隔离
- 任务开始时:调用 mark_processing() 更新为 running
- 任务成功时:调用 mark_completed() 更新为 completed,传入视频数量
- 任务失败时:在 except 块中调用 mark_failed() 记录错误信息
- _create_video_record_and_dedup 返回值改为 int,方便统计成功数量

状态更新使用独立数据库 session,且所有状态更新操作都包裹在 try/except 中,确保不会因为状态更新失败导致整个任务异常。

### 3. 新增状态机单元测试
新增 tests/unit/test_generation_task_status.py,覆盖 42 个测试用例:
- 初始状态验证
- mark_processing 正常/异常路径
- mark_completed 正常/异常路径
- mark_failed 正常/异常路径
- mark_cancelled 正常/异常路径
- mark_pending_from_failed 重试路径
- transition_to 通用方法
- 完整流转路径(成功、失败、重试、取消)

## 影响范围
- packages/domain/generation_task.py — 领域模型状态机方法
- apps/worker/worker_app/tasks/generation.py — Celery 任务状态更新
- tests/unit/test_generation_task_status.py — 新增单元测试

## 验证方式
- 42 个状态机单元测试全部通过
- 现有 8 个生成相关测试全部通过
- 现有 15 个生成 API 集成测试全部通过
- 现有 6 个编辑计划 worker 失败测试全部通过
- 现有 25 个生成视频管理测试全部通过
2026-07-09 15:34:35 +08:00

248 lines
8.0 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""GenerationTask 领域模型 — 视频生成任务.
状态机:
pending → running → completed
↘ failed → pending (重试)
↘ cancelled
"""
from __future__ import annotations
import sys
from dataclasses import dataclass, field
from datetime import datetime, timezone
if sys.version_info >= (3, 11):
from enum import StrEnum
else:
from enum import Enum
class StrEnum(str, Enum):
pass
from uuid import uuid4
class GenerationTaskStatus(StrEnum):
"""生成任务状态枚举。"""
PENDING = "pending"
"""待处理(任务已创建,等待执行)"""
RUNNING = "running"
"""运行中(正在生成视频)"""
COMPLETED = "completed"
"""已完成(视频生成成功)"""
FAILED = "failed"
"""失败(生成失败)"""
CANCELLED = "cancelled"
"""已取消(用户取消或系统取消)"""
# 终态集合
TERMINAL_STATUSES = frozenset(
{GenerationTaskStatus.COMPLETED, GenerationTaskStatus.FAILED, GenerationTaskStatus.CANCELLED}
)
# 合法状态转换
_VALID_TRANSITIONS: dict[GenerationTaskStatus, set[GenerationTaskStatus]] = {
GenerationTaskStatus.PENDING: {
GenerationTaskStatus.RUNNING,
GenerationTaskStatus.FAILED,
GenerationTaskStatus.CANCELLED,
},
GenerationTaskStatus.RUNNING: {
GenerationTaskStatus.COMPLETED,
GenerationTaskStatus.FAILED,
GenerationTaskStatus.CANCELLED,
},
GenerationTaskStatus.FAILED: {GenerationTaskStatus.PENDING}, # 重试回到 pending
}
@dataclass(slots=True)
class GenerationTask:
id: str
project_id: str
asset_library_id: str
strategy_id: str = ""
voice_library_id: str = ""
template_id: str = ""
asset_ids: list[str] = field(default_factory=list)
title_ids: list[str] = field(default_factory=list)
voice_ids: list[str] = field(default_factory=list)
status: GenerationTaskStatus = GenerationTaskStatus.PENDING
progress: float = 0.0
result_count: int = 0
error_message: str = ""
started_at: datetime | None = None
completed_at: datetime | None = None
source_edit_plan_id: str = ""
created_by_user_id: str = ""
asset_select_mode: str = ""
batch_id: str = ""
created_at: datetime = field(default_factory=lambda: datetime.now(timezone.utc))
@classmethod
def create(
cls,
project_id: str,
asset_library_id: str,
*,
strategy_id: str = "",
voice_library_id: str = "",
template_id: str = "",
asset_ids: list[str] | None = None,
title_ids: list[str] | None = None,
voice_ids: list[str] | None = None,
created_by_user_id: str = "",
source_edit_plan_id: str = "",
asset_select_mode: str = "",
batch_id: str = "",
) -> "GenerationTask":
if not project_id.strip() and not template_id.strip():
raise ValueError("project_id 或 template_id 至少需要提供一个")
if not asset_library_id.strip() and not (asset_ids or title_ids or voice_ids):
raise ValueError("asset_library_id 或 asset_ids/title_ids/voice_ids 至少需要提供一个")
return cls(
id=uuid4().hex,
project_id=project_id.strip(),
asset_library_id=asset_library_id.strip(),
strategy_id=strategy_id.strip(),
voice_library_id=voice_library_id.strip(),
template_id=template_id.strip(),
asset_ids=list(asset_ids) if asset_ids else [],
title_ids=list(title_ids) if title_ids else [],
voice_ids=list(voice_ids) if voice_ids else [],
created_by_user_id=created_by_user_id.strip(),
source_edit_plan_id=source_edit_plan_id.strip(),
asset_select_mode=asset_select_mode,
batch_id=batch_id,
)
# ── 状态查询 ────────────────────────────────────────────────────────────
@property
def is_terminal(self) -> bool:
"""是否处于终态(completed / failed / cancelled)。"""
return self.status in TERMINAL_STATUSES
@property
def is_completed(self) -> bool:
"""是否已完成。"""
return self.status == GenerationTaskStatus.COMPLETED
@property
def is_failed(self) -> bool:
"""是否失败。"""
return self.status == GenerationTaskStatus.FAILED
@property
def is_running(self) -> bool:
"""是否运行中。"""
return self.status == GenerationTaskStatus.RUNNING
# ── 状态转换 ────────────────────────────────────────────────────────────
def transition_to(self, new_status: GenerationTaskStatus | str) -> None:
"""执行状态转换。
Args:
new_status: 目标状态
Raises:
ValueError: 非法状态转换
"""
if isinstance(new_status, str):
try:
new_status = GenerationTaskStatus(new_status)
except ValueError:
raise ValueError(f"无效状态: {new_status}")
allowed = _VALID_TRANSITIONS.get(self.status, set())
if new_status not in allowed:
raise ValueError(
f"非法状态转换: {self.status.value} → {new_status.value},"
f"允许: {{{', '.join(sorted(s.value for s in allowed))}}}"
)
self.status = new_status
def mark_processing(self) -> None:
"""标记为处理中(pending → running)。
设置 started_at,清除 error_message。
Raises:
ValueError: 当前状态不允许转换到 running
"""
self.transition_to(GenerationTaskStatus.RUNNING)
self.started_at = datetime.now(timezone.utc)
self.error_message = ""
def mark_completed(self, result_count: int = 1) -> None:
"""标记为已完成(running → completed)。
设置 completed_at、progress=100.0、result_count,清除 error_message。
Args:
result_count: 生成的视频数量,默认为 1
Raises:
ValueError: 当前状态不允许转换到 completed
"""
self.transition_to(GenerationTaskStatus.COMPLETED)
self.completed_at = datetime.now(timezone.utc)
self.progress = 100.0
self.result_count = result_count
self.error_message = ""
def mark_failed(self, error_message: str) -> None:
"""标记为失败(pending / running → failed)。
设置 error_message、completed_at。
Args:
error_message: 错误信息
Raises:
ValueError: 当前状态不允许转换到 failed
"""
self.transition_to(GenerationTaskStatus.FAILED)
self.error_message = error_message
self.completed_at = datetime.now(timezone.utc)
def mark_cancelled(self) -> None:
"""标记为已取消(pending / running → cancelled)。
设置 completed_at。
Raises:
ValueError: 当前状态不允许转换到 cancelled
"""
self.transition_to(GenerationTaskStatus.CANCELLED)
self.completed_at = datetime.now(timezone.utc)
def mark_pending_from_failed(self) -> None:
"""从失败状态重置为待处理(用于重试)。
清除 error_message、started_at、completed_at、progress。
Raises:
ValueError: 当前状态不是 failed
"""
if self.status != GenerationTaskStatus.FAILED:
raise ValueError(
f"只有 failed 状态的任务可以重置为 pending,当前状态: {self.status.value}"
)
self.transition_to(GenerationTaskStatus.PENDING)
self.error_message = ""
self.started_at = None
self.completed_at = None
self.progress = 0.0
self.result_count = 0