Files
intelligence_system/system_management/scheduler/task_scheduler.py
T
2025-08-07 17:58:41 +08:00

277 lines
9.7 KiB
Python

# system_management/scheduler/task_scheduler.py
import importlib
import time
from datetime import datetime, timedelta
from typing import Dict, List, Optional
from storage.mysql_agent import MySQLAgent
from pathlib import Path
# 使用您的日志系统
from utils.logger import CrossPlatformLog
log = CrossPlatformLog.get_logger("TaskScheduler")
class TaskScheduler:
def __init__(self, db_config: Optional[Dict] = None):
"""
初始化任务调度器
Args:
db_config (Optional[Dict]): 可选的数据库配置,默认使用MySQLAgent默认配置
"""
self.db = MySQLAgent(db_config or {}) # 使用您提供的MySQLAgent
self._init_task_table()
log.info("TaskScheduler initialized")
def _init_task_table(self):
"""确保任务表存在并包含必要字段"""
if not self.db.table_exists("main_task"):
log.info("Creating main_task table")
create_sql = """
CREATE TABLE main_task (
task_id INT AUTO_INCREMENT PRIMARY KEY,
task_name VARCHAR(100) NOT NULL,
module_path VARCHAR(255) NOT NULL COMMENT '例如data_collection.spiders.weibo_spider',
frequency_type ENUM('minute','hourly','daily','weekly','monthly') NOT NULL,
frequency_value INT DEFAULT NULL COMMENT '间隔数值',
last_run_time DATETIME DEFAULT NULL,
next_run_time DATETIME DEFAULT NULL,
last_run_status VARCHAR(20) DEFAULT NULL,
is_active TINYINT(1) DEFAULT 1,
is_running TINYINT(1) DEFAULT 0,
run_count INT DEFAULT 0,
created_at DATETIME DEFAULT CURRENT_TIMESTAMP,
updated_at DATETIME DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
INDEX idx_next_run (next_run_time),
INDEX idx_active (is_active)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4
"""
self.db.execute_sql(create_sql)
log.success("main_task table created")
def run_pending_tasks(self) -> Dict[str, int]:
"""
执行所有到期的活跃任务
Returns:
Dict[str, int]: 包含执行结果的字典 {
'total': 总任务数,
'success': 成功数,
'failed': 失败数
}
"""
result = {'total': 0, 'success': 0, 'failed': 0}
try:
# 使用您提供的query_to_df方法获取任务
tasks_df = self.db.query_to_df(
"SELECT * FROM main_task "
"WHERE is_active = 1 AND next_run_time <= %s "
"ORDER BY next_run_time",
params=(datetime.now(),)
)
result['total'] = len(tasks_df)
if tasks_df.empty:
log.debug("No pending tasks found")
return result
for _, task in tasks_df.iterrows():
task_id = task['task_id']
log.bind(task_id=task_id).info(
f"Starting task {task['task_name']}"
)
# 标记任务为执行中
self._update_task_status(
task_id,
{'is_running': 1, 'last_run_time': datetime.now()}
)
try:
self._execute_single_task(task)
self._update_task_status(
task_id,
{
'last_run_status': 'success',
'is_running': 0,
'run_count': task['run_count'] + 1,
'next_run_time': self._calculate_next_run(
task['frequency_type'],
task['frequency_value']
)
}
)
result['success'] += 1
log.bind(task_id=task_id).success("Task completed")
except Exception as e:
log.bind(task_id=task_id).error(
f"Task failed: {str(e)}",
exc_info=True
)
self._update_task_status(
task_id,
{
'last_run_status': 'failed',
'is_running': 0,
'next_run_time': self._calculate_next_run(
task['frequency_type'],
task['frequency_value'],
retry=True
)
}
)
result['failed'] += 1
log.info(
"Scheduler cycle completed",
total_tasks=result['total'],
success=result['success'],
failed=result['failed']
)
return result
except Exception as e:
log.critical(
"Scheduler main loop failed",
exc_info=True
)
raise
def _execute_single_task(self, task: Dict) -> None:
"""执行单个任务模块"""
start_time = time.time()
task_log = log.bind(
task_id=task['task_id'],
module=task['module_path']
)
try:
module = importlib.import_module(task['module_path'])
if not hasattr(module, 'main'):
raise ImportError(f"Module has no main() function")
# 执行任务
task_log.debug("Task execution started")
module.main()
elapsed = time.time() - start_time
task_log.info(
f"Task completed in {elapsed:.2f}s",
duration=elapsed
)
except Exception as e:
task_log.error(
"Task execution failed",
exc_info=True
)
raise
def _update_task_status(self, task_id: int, updates: Dict) -> None:
"""更新任务状态"""
set_clause = ", ".join([f"{k}=%s" for k in updates.keys()])
sql = f"UPDATE main_task SET {set_clause} WHERE task_id=%s"
params = list(updates.values()) + [task_id]
try:
affected = self.db.execute_sql(sql, params=params)
if affected != 1:
log.warning(
"Unexpected row count in update",
task_id=task_id,
expected=1,
affected=affected
)
except Exception as e:
log.error(
"Failed to update task status",
task_id=task_id,
exc_info=True
)
raise
def _calculate_next_run(self, freq_type: str, freq_value: Optional[int] = None,
retry: bool = False) -> datetime:
"""
计算下次执行时间(带重试逻辑)
"""
base_time = datetime.now()
if retry:
# 失败后15分钟重试
log.debug("Calculating retry time")
return base_time + timedelta(minutes=15)
if freq_type == 'minute':
delta = timedelta(minutes=freq_value or 1)
elif freq_type == 'hourly':
delta = timedelta(hours=freq_value or 1)
elif freq_type == 'daily':
delta = timedelta(days=freq_value or 1)
elif freq_type == 'weekly':
delta = timedelta(weeks=freq_value or 1)
elif freq_type == 'monthly':
# 处理月末日期特殊情况
next_month = (base_time.replace(day=1) + timedelta(days=32)).replace(day=1)
last_day = (next_month - timedelta(days=1)).day
day = min(base_time.day, last_day)
return base_time.replace(day=1, month=next_month.month, day=day)
else:
raise ValueError(f"Unknown frequency type: {freq_type}")
return base_time + delta
def add_task(self, task_name: str, module_path: str, frequency_type: str,
frequency_value: Optional[int] = None) -> int:
"""
添加新任务到调度系统
"""
sql = """
INSERT INTO main_task
(task_name, module_path, frequency_type, frequency_value, next_run_time)
VALUES (%s, %s, %s, %s, %s)
"""
next_run = self._calculate_next_run(frequency_type, frequency_value)
params = (task_name, module_path, frequency_type, frequency_value, next_run)
try:
self.db.execute_sql(sql, params=params)
task_id = self.db.query_to_df("SELECT LAST_INSERT_ID() AS id").iloc[0]['id']
log.info(
"New task added",
task_id=task_id,
task_name=task_name,
next_run=next_run
)
return task_id
except Exception as e:
log.error(
"Failed to add new task",
task_name=task_name,
exc_info=True
)
raise
def get_task_status(self, active_only: bool = True) -> pd.DataFrame:
"""
获取任务状态
"""
where = "WHERE is_active = 1" if active_only else ""
log.debug("Fetching task status", active_only=active_only)
return self.db.query_to_df(
f"""
SELECT
task_id, task_name, module_path,
frequency_type, frequency_value,
last_run_time, next_run_time,
last_run_status, run_count,
is_active, is_running
FROM main_task
{where}
ORDER BY next_run_time
"""
)