auto_news_scheduler/utils_timing_remove_data.py

# -*- coding: utf-8 -*-
"""
自动清除大于指定天数的数据
"""
import threading
import time
import sys
import os
from datetime import datetime
import pymongo
import smtplib
from email.mime.text import MIMEText
from email.header import Header

sys.path.append(os.path.join(os.getcwd().split('auto')[0], 'auto'))
base_project = os.path.join(os.getcwd().split('auto')[0], 'auto')
import json

config_path = os.path.join(base_project, 'config.json')
config_json = {}
with open(config_path, 'r') as f:
    config_json = json.load(f)

if not config_json:
    print('No config file found')
    exit(0)

PROJECT_NAME = config_json.get('PROJECT_NAME')
DB_USER = config_json.get('DB_USER')
DB_PASSWORD = config_json.get('DB_PASSWORD')
DB_IP = config_json.get('DB_IP')
DB_PORT = config_json.get('DB_PORT')
MONGO_LINK = f'mongodb://{DB_USER}:{DB_PASSWORD}@{DB_IP}:{DB_PORT}/'
MAIL_HOST = config_json.get('MAIL_HOST')
MAIL_USER = config_json.get('MAIL_USER')
MAIL_PASS = config_json.get('MAIL_PASS')
MAIL_SENDER = config_json.get('MAIL_SENDER')
MAIL_RECEIVERS = config_json.get('MAIL_RECEIVERS')

now_day = time.strftime('%Y-%m-%d', time.localtime())
rss_base_url = 'http://home.erhe.link:20002/xmlfile/'
base_project = os.path.join(os.getcwd().split(PROJECT_NAME)[0], PROJECT_NAME)


class MongoHandle(object):
    def __init__(self, db, collection, del_db=False, del_collection=False, auto_remove=0):
        self.client = pymongo.MongoClient(MONGO_LINK)
        self.db = db
        self.collection = collection

        if del_db and db:
            # 检查数据库是否存在
            if db in self.client.list_database_names():
                # 删除数据库
                self.client.drop_database(db)
        self.db = self.client[db]

        if del_collection and self.collection:
            # 检查集合是否存在
            if self.collection in self.db.list_collection_names():
                # 删除集合
                self.db.drop_collection(collection)
        self.collection = self.db[collection]

        if auto_remove:
            self.auto_remove_data(auto_remove)

    def write_data(self, data):
        self.collection.insert_one(data)

    def auto_remove_data(self, day):
        for data in self.collection.find({'create_time': {'$lt': int(time.time()) - day * 24 * 60 * 60}}):
            self.collection.delete_one({'_id': data['_id']})


class SendEmail(object):
    def __init__(self, subject='auto subject', title='auto title', text='auto text') -> None:
        # 第三方 SMTP 服务
        self.mail_host = MAIL_HOST  # 设置服务器
        self.mail_user = MAIL_USER  # 用户名
        self.mail_pass = MAIL_PASS  # 口令

        self.sender = MAIL_SENDER
        self.receivers = [MAIL_RECEIVERS]

        self.subject = subject
        self.title = title
        self.text = text

    def send(self):
        message = MIMEText(self.text, 'plain', 'utf-8')
        message['From'] = Header(self.title, 'utf-8')
        message['To'] = Header("auto collection", 'utf-8')

        subject = self.subject
        message['Subject'] = Header(subject, 'utf-8')

        try:
            smtpObj = smtplib.SMTP()
            smtpObj.connect(self.mail_host, 25)
            smtpObj.login(self.mail_user, self.mail_pass)
            smtpObj.sendmail(self.sender, self.receivers, message.as_string())
            print("邮件发送成功")
        except smtplib.SMTPException:
            print("Error: 无法发送邮件")


class LogsHandle(object):
    def __init__(self):
        self.now_day = time.strftime('%Y-%m-%d', time.localtime())
        db = 'logs'
        collection = 'logs_' + self.now_day
        self.mongo = MongoHandle(db=db, collection=collection, del_db=False, del_collection=False, auto_remove=0)

    def logs_write(self, title_source=None, content=None, state=None, send_now=False):
        data_to_insert = {
            "title": title_source,
            "context": content,
            "state": state,
            "create_time": int(time.time()),
            "create_datetime": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
        }

        self.mongo.collection.insert_one(data_to_insert)

        if send_now:
            subject = 'auto collection'
            title = 'auto collection - running logs: {}'.format(self.now_day)
            text = 'logs_source: {}, logs_detail: {}, state: {} logs_create_time: {}'.format(data_to_insert.setdefault('title'),
                                                                                             data_to_insert.setdefault('content'),
                                                                                             data_to_insert.setdefault('state'),
                                                                                             data_to_insert.setdefault('create_datetime'),
                                                                                             )

            Send = SendEmail(subject=subject, title=title, text=text)
            Send.send()


class AutoRemoveData(object):
    def __init__(self):
        self.databases = [
            'news',
            'apprcn',
            'HelloGithub'
        ]

        self.day = 60

        self.client = pymongo.MongoClient(MONGO_LINK)

        self.logs = LogsHandle()

        self.all_delete_count = 0

    def auto_remove_data(self, db_name, day):
        print(f'准备删除时间大于: {self.day} 数据')
        if db_name not in self.client.list_database_names():
            return

        deleted_count = 0

        db = self.client[db_name]

        for collection_name in db.list_collection_names():
            collection = db[collection_name]
            for data in collection.find({'create_time': {'$lt': int(time.time()) - day * 24 * 60 * 60}}):
                collection.delete_one({'_id': data['_id']})
                deleted_count += 1

        self.all_delete_count += deleted_count

        msg = f"删除 {db_name} 库 {self.day} 天以上数据 {deleted_count} 条"
        if deleted_count:
            print(msg)
            self.logs.logs_write(f'自动删除 {self.day} 天以上数据', msg, 'delete', False)

    def main(self):
        self.logs.logs_write(f'自动删除 {self.day} 天以上数据', f'开始自动删除 {self.day} 天以上数据', 'start', False)
        threads = []

        for db_name in self.databases:
            thread = threading.Thread(target=self.auto_remove_data, args=(db_name, self.day))
            threads.append(thread)
            thread.start()

        for thread in threads:
            thread.join()

        print(f'删除时间大于: {self.day} 数据, 已完成')
        print(f'本次运行共删除: {self.all_delete_count} 条数据')
        self.logs.logs_write(f'自动删除 {self.day} 天以上数据', f'自动删除 {self.day} 天数以上数据完成', 'done', False)


if __name__ == "__main__":
    A = AutoRemoveData()
    A.main()