利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Python ai-video-generatorcontent-creationffmpeginstagram-reelsllmpythonshort-videosubtitlestext-to-speechtiktokvideo-automationvideo-workflowworkflow-automationyoutube-shorts
fix(worker): handle stale queued task params
kapil971390254cd028906e
diff --git a/app/controllers/manager/base_manager.py b/app/controllers/manager/base_manager.pyindex ac5bf710..ab214643 100644--- a/app/controllers/manager/base_manager.py+++ b/app/controllers/manager/base_manager.py@@ -70,6 +70,12 @@ class TaskManager:and not self.is_queue_empty()):task_info = self.dequeue()+ if task_info is None:+ # dequeue() may skip and discard queue entries that no longer+ # pass current validation (see RedisTaskManager.dequeue) and+ # return None once nothing usable is left, even though+ # is_queue_empty() was False a moment earlier.+ returnfunc = task_info["func"]args = task_info.get("args", ())kwargs = task_info.get("kwargs", {})diff --git a/app/controllers/manager/redis_manager.py b/app/controllers/manager/redis_manager.pyindex db108cc2..2f9568aa 100644--- a/app/controllers/manager/redis_manager.py+++ b/app/controllers/manager/redis_manager.py@@ -2,9 +2,13 @@ import jsonfrom typing import Dictimport redis+from loguru import logger+from pydantic import ValidationErrorfrom app.controllers.manager.base_manager import TaskManager+from app.models import constfrom app.models.schema import VideoParams+from app.services import state as smfrom app.services import task as tmFUNC_MAP = {@@ -44,8 +48,17 @@ class RedisTaskManager(TaskManager):self.redis_client.rpush(self.queue, json.dumps(task_with_serializable_params))def dequeue(self):- task_json = self.redis_client.lpop(self.queue)- if task_json:+ # 循环而非单次弹出:某个任务在入队时可能满足当时的 VideoParams 校验规则,+ # 但校验规则本身在两次部署之间收紧了(例如新增 ge=1 约束)。lpop 是破坏性+ # 操作,一旦弹出就不能放回原位;如果重建 VideoParams 时才发现校验失败,+ # 这条任务已经从队列中永久移除了,不能再假装它还在。与其让异常从这里往上+ # 抛、把这条已经丢失的任务的 lock 持有者带崩,不如原地丢弃并继续尝试队列+ # 里的下一条,把"拿到一条可用任务或者队列确实空了"这个约定维持住。+ while True:+ task_json = self.redis_client.lpop(self.queue)+ if not task_json:+ return None+task_info = json.loads(task_json)# 将函数名称转换回函数对象task_info["func"] = FUNC_MAP[task_info["func"]]@@ -53,12 +66,31 @@ class RedisTaskManager(TaskManager):if "params" in task_info["kwargs"] and isinstance(task_info["kwargs"]["params"], dict):- task_info["kwargs"]["params"] = VideoParams(- **task_info["kwargs"]["params"]- )+ try:+ task_info["kwargs"]["params"] = VideoParams(+ **task_info["kwargs"]["params"]+ )+ except ValidationError as e:+ logger.error(+ "dropping queued task with params that fail current "+ f"VideoParams validation (queued under an older, more "+ f"permissive schema, or corrupted): {e}"+ )+ # 任务状态记录在入队前就已创建,且默认是 processing;如果只是+ # 丢弃这条队列项而不动状态记录,API/WebUI 会一直显示任务在+ # 运行,永远不会变成失败。用 patch_task 而不是 update_task,+ # 这样如果用户已经删除了这个任务,我们不会又把它建回来。+ task_id = task_info["kwargs"].get("task_id")+ if task_id:+ sm.state.patch_task(+ task_id,+ state=const.TASK_STATE_FAILED,+ failed_stage="dequeue",+ error=f"discarded stale queued task: {e}",+ )+ continuereturn task_info- return Nonedef is_queue_empty(self):return self.redis_client.llen(self.queue) == 0diff --git a/test/services/test_task_manager.py b/test/services/test_task_manager.pyindex 428ae02a..3ec25594 100644--- a/test/services/test_task_manager.py+++ b/test/services/test_task_manager.py@@ -5,6 +5,7 @@ from unittest.mock import MagicMock, patchfrom app.controllers.manager.base_manager import TaskQueueFullErrorfrom app.controllers.manager.memory_manager import InMemoryTaskManagerfrom app.controllers.manager.redis_manager import RedisTaskManager+from app.models import constfrom app.models.schema import VideoParamsfrom app.services import task as task_service@@ -103,6 +104,22 @@ class TestInMemoryTaskManager(unittest.TestCase):self.assertEqual(manager.current_tasks, 1)task_done.assert_called_once_with()+ def test_check_queue_handles_dequeue_returning_none(self):+ """+ dequeue() 可能在内部跳过所有已不满足当前校验的排队任务后返回 None,+ 即使调用 check_queue 之前 is_queue_empty() 曾经是 False。check_queue+ 不能假设 dequeue 一定能拿到可用任务,否则会在 task_info["func"] 上崩溃。+ """+ manager = InMemoryTaskManager(max_concurrent_tasks=1, max_queued_tasks=1)++ with patch.object(manager, "is_queue_empty", return_value=False), patch.object(+ manager, "dequeue", return_value=None+ ), patch.object(manager, "execute_task") as execute_task:+ manager.check_queue()++ execute_task.assert_not_called()+ self.assertEqual(manager.current_tasks, 0)+def test_execute_task_starts_background_thread(self):"""任务执行入口必须启动线程,并把函数参数完整传给 run_task。"""manager = InMemoryTaskManager(max_concurrent_tasks=1)@@ -190,6 +207,116 @@ class TestRedisTaskManager(unittest.TestCase):self.assertTrue(self.manager.is_queue_empty())self.assertEqual(self.manager.queue_size(), 2)+ def test_dequeue_skips_task_that_fails_current_validation(self):+ """+ 一条任务可能是在校验规则收紧前入队的(例如 video_count 曾允许为 0)。+ lpop 是破坏性操作,重建 VideoParams 失败时这条任务已经从 Redis 里+ 永久移除了,不能再假装它还在;dequeue 不应该把校验异常抛给调用方+ (那样会让持锁的调用方崩溃且丢失这条任务却不打日志),而应该跳过它,+ 继续尝试队列里的下一条,直到取到一条可用任务或者队列确实空了。+ """+ stale_payload = {+ "func": "start",+ "args": [],+ "kwargs": {+ "task_id": "task-stale",+ "params": {**VideoParams(video_subject="Coffee").model_dump(+ warnings=False+ ), "video_count": 0},+ },+ }+ valid_payload = {+ "func": "start",+ "args": [],+ "kwargs": {+ "task_id": "task-valid",+ "params": VideoParams(video_subject="Tea").model_dump(+ warnings=False+ ),+ },+ }+ self.redis_client.lpop.side_effect = [+ json.dumps(stale_payload),+ json.dumps(valid_payload),+ ]++ task = self.manager.dequeue()++ self.assertEqual(self.redis_client.lpop.call_count, 2)+ self.assertEqual(task["kwargs"]["task_id"], "task-valid")+ self.assertIsInstance(task["kwargs"]["params"], VideoParams)+ self.assertEqual(task["kwargs"]["params"].video_subject, "Tea")++ def test_dequeue_returns_none_when_every_queued_task_is_stale(self):+ """全部剩余任务都因当前校验规则被丢弃时,应返回 None 而不是抛出异常。"""+ stale_payload = {+ "func": "start",+ "args": [],+ "kwargs": {+ "task_id": "task-stale",+ "params": {**VideoParams(video_subject="Coffee").model_dump(+ warnings=False+ ), "video_count": -1},+ },+ }+ self.redis_client.lpop.side_effect = [json.dumps(stale_payload), None]++ self.assertIsNone(self.manager.dequeue())+ self.assertEqual(self.redis_client.lpop.call_count, 2)++ def test_dequeue_marks_stale_task_failed_instead_of_leaving_it_processing(self):+ """+ 任务状态记录在入队前就已创建,默认是 processing。仅仅在 dequeue 里跳过+ 并丢弃这条队列项而不更新状态记录,会让这个任务在 API/WebUI 里永远显示+ 为运行中。应该用 patch_task(而不是 update_task)把它标记为失败,+ 这样如果任务已经被用户删除,我们不会又把它的状态记录建回来。+ """+ stale_payload = {+ "func": "start",+ "args": [],+ "kwargs": {+ "task_id": "task-stale",+ "params": {**VideoParams(video_subject="Coffee").model_dump(+ warnings=False+ ), "video_count": 0},+ },+ }+ self.redis_client.lpop.side_effect = [json.dumps(stale_payload), None]++ with patch("app.controllers.manager.redis_manager.sm.state") as state:+ state.patch_task.return_value = True+ result = self.manager.dequeue()++ self.assertIsNone(result)+ state.patch_task.assert_called_once()+ call_args = state.patch_task.call_args+ self.assertEqual(call_args.args[0], "task-stale")+ self.assertEqual(call_args.kwargs["state"], const.TASK_STATE_FAILED)+ self.assertEqual(call_args.kwargs["failed_stage"], "dequeue")+ self.assertIn("video_count", call_args.kwargs["error"])++ def test_dequeue_does_not_recreate_state_for_already_deleted_task(self):+ """patch_task 在任务已被删除时返回 False;dequeue 不应把它当成错误处理。"""+ stale_payload = {+ "func": "start",+ "args": [],+ "kwargs": {+ "task_id": "task-deleted",+ "params": {**VideoParams(video_subject="Coffee").model_dump(+ warnings=False+ ), "video_count": 0},+ },+ }+ self.redis_client.lpop.side_effect = [json.dumps(stale_payload), None]++ with patch("app.controllers.manager.redis_manager.sm.state") as state:+ state.patch_task.return_value = False+ result = self.manager.dequeue()++ self.assertIsNone(result)+ state.patch_task.assert_called_once()+ state.update_task.assert_not_called()+if __name__ == "__main__":unittest.main()