zhouxiaoka--autoclip
wehub-resource-sync
149 行
5.0 KiB
Python
149 行
5.0 KiB
Python
"""
|
||
任务提交工具
|
||
独立的工具函数，避免循环导入问题
|
||
"""
|
||
|
||
import logging
|
||
import os
|
||
from typing import Dict, Any, Optional
|
||
from ..core.celery_app import celery_app
|
||
|
||
logger = logging.getLogger(__name__)
|
||
|
||
|
||
def _is_desktop_mode() -> bool:
|
||
return os.getenv("AUTOCLIP_DESKTOP_MODE", "").lower() in {"1", "true", "yes"}
|
||
|
||
|
||
def _run_pipeline_locally(project_id: str, input_video_path: str, input_srt_path: str) -> Dict[str, Any]:
|
||
"""桌面模式：不经过 Redis/Celery broker，直接在后台线程内同步执行流水线任务。
|
||
|
||
桌面安装包里没有 Redis，而生产用的 core.celery_app 走 redis://localhost。
|
||
Celery 任务 process_video_pipeline 本身是「在任务内同步跑完整条流水线」
|
||
（asyncio.run(pipeline_adapter...)），不会再派发子任务，所以可以用 .apply()
|
||
在本地线程里直接跑，进度照常写进数据库的 Task 记录供前端轮询。
|
||
"""
|
||
import uuid
|
||
import threading
|
||
|
||
task_id = str(uuid.uuid4())
|
||
|
||
def run():
|
||
try:
|
||
# 延迟导入，避免循环依赖
|
||
from ..tasks.processing import process_video_pipeline
|
||
process_video_pipeline.apply(
|
||
args=[project_id, input_video_path, input_srt_path],
|
||
task_id=task_id,
|
||
)
|
||
logger.info(f"桌面模式本地流水线执行结束: {project_id}, task_id={task_id}")
|
||
except Exception as e: # noqa: BLE001
|
||
logger.error(f"桌面模式本地流水线执行失败: {project_id}, 错误: {e}", exc_info=True)
|
||
|
||
threading.Thread(target=run, name=f"pipeline-{project_id[:8]}", daemon=True).start()
|
||
logger.info(f"桌面模式：已在本地后台线程启动视频流水线 {project_id}, task_id={task_id}")
|
||
return {
|
||
'success': True,
|
||
'task_id': task_id,
|
||
'status': 'PENDING',
|
||
'message': '视频流水线任务已在本地启动',
|
||
}
|
||
|
||
|
||
def submit_video_pipeline_task(project_id: str, input_video_path: str, input_srt_path: str) -> Dict[str, Any]:
|
||
"""
|
||
提交视频流水线任务
|
||
|
||
Args:
|
||
project_id: 项目ID
|
||
input_video_path: 输入视频路径
|
||
input_srt_path: 输入SRT路径
|
||
|
||
Returns:
|
||
任务提交结果
|
||
"""
|
||
# 桌面模式没有 Redis，走本地线程执行
|
||
if _is_desktop_mode():
|
||
return _run_pipeline_locally(project_id, input_video_path, input_srt_path)
|
||
|
||
try:
|
||
logger.info(f"提交视频流水线任务: {project_id}")
|
||
|
||
# 直接使用celery_app提交任务
|
||
logger.info(f"准备提交任务到队列...")
|
||
logger.info(f"任务名称: backend.tasks.processing.process_video_pipeline")
|
||
logger.info(f"任务参数: {[project_id, input_video_path, input_srt_path]}")
|
||
|
||
try:
|
||
celery_task = celery_app.send_task(
|
||
'backend.tasks.processing.process_video_pipeline',
|
||
args=[project_id, input_video_path, input_srt_path]
|
||
)
|
||
|
||
logger.info(f"视频流水线任务已提交: {celery_task.id}")
|
||
logger.info(f"任务状态: {celery_task.state}")
|
||
|
||
# 检查任务是否真的提交到队列
|
||
import redis
|
||
r = redis.Redis(host='localhost', port=6379, db=0)
|
||
queue_length = r.llen('processing')
|
||
logger.info(f"Redis队列长度: {queue_length}")
|
||
|
||
except Exception as e:
|
||
logger.error(f"任务提交过程中出现异常: {e}")
|
||
raise
|
||
|
||
return {
|
||
'success': True,
|
||
'task_id': celery_task.id,
|
||
'status': 'PENDING',
|
||
'message': '视频流水线任务已提交'
|
||
}
|
||
|
||
except Exception as e:
|
||
logger.error(f"提交视频流水线任务失败: {project_id}, 错误: {e}")
|
||
return {
|
||
'success': False,
|
||
'error': str(e),
|
||
'message': '任务提交失败'
|
||
}
|
||
|
||
def submit_single_step_task(project_id: str, step: str, config: Dict[str, Any]) -> Dict[str, Any]:
|
||
"""
|
||
提交单个步骤任务
|
||
|
||
Args:
|
||
project_id: 项目ID
|
||
step: 步骤名称
|
||
config: 处理配置
|
||
|
||
Returns:
|
||
任务提交结果
|
||
"""
|
||
try:
|
||
logger.info(f"提交单个步骤任务: {project_id}, {step}")
|
||
|
||
# 直接使用celery_app提交任务
|
||
celery_task = celery_app.send_task(
|
||
'tasks.processing.process_single_step',
|
||
args=[project_id, step, config]
|
||
)
|
||
|
||
logger.info(f"单个步骤任务已提交: {celery_task.id}")
|
||
|
||
return {
|
||
'success': True,
|
||
'task_id': celery_task.id,
|
||
'step': step,
|
||
'status': 'PENDING',
|
||
'message': f'步骤 {step} 任务已提交'
|
||
}
|
||
|
||
except Exception as e:
|
||
logger.error(f"提交单个步骤任务失败: {project_id}, {step}, 错误: {e}")
|
||
return {
|
||
'success': False,
|
||
'error': str(e),
|
||
'message': '任务提交失败'
|
||
}
|