piaoquan_crawler/dev/dev_script/get_intervals.py

# -*- coding: utf-8 -*-
# @Author: wangkun
# @Time: 2023/7/18
import datetime

# 读取日志文件并将每行日志存储到一个列表中
# with open('../logs/benshanzhufu-recommend-2023-07-17.log', 'r') as file:
# with open('../logs/douyin-author-2023-07-17.log', 'r') as file:
with open('../logs/douyin-recommend-2023-07-19.log', 'r') as file:
# with open('../logs/kuaishou-author-2023-07-17.log', 'r') as file:
# with open('../logs/kuaishou-recommend-2023-07-17.log', 'r') as file:
# with open('../logs/xigua-author-2023-07-17.log', 'r') as file:
# with open('../logs/xigua-search-2023-07-17.log', 'r') as file:
# with open('../logs/xiaoniangao-author-2023-07-17.log', 'r') as file:
# with open('../logs/xiaoniangao-play-2023-07-17.log', 'r') as file:
# with open('../logs/xiaoniangao-hour-2023-07-17.log', 'r') as file:
# with open('../logs/suisuiniannianyingfuqi-recommend-2023-07-17.log', 'r') as file:
# with open('../logs/gongzhonghao-author1-2023-07-18.log', 'r') as file:
# with open('../logs/gongzhonghao-author2-2023-07-18.log', 'r') as file:
# with open('../logs/gongzhonghao-author3-2023-07-18.log', 'r') as file:
# with open('../logs/gongzhonghao-author4-2023-07-18.log', 'r') as file:
# with open('../logs/gongzhonghao-author5-2023-07-18.log', 'r') as file:
# with open('../logs/gongzhonghao-author6-2023-07-18.log', 'r') as file:
    log_lines = file.readlines()

# 存储间隔时间的列表
intervals = []

# 遍历日志列表，计算相邻两条日志的时间间隔
for i in range(1, len(log_lines)):
    if "2023-" not in log_lines[i-1] or "2023-" not in log_lines[i]:
        continue
    # 解析时间戳
    timestamp1 = datetime.datetime.strptime(log_lines[i - 1].split(".")[0], '%Y-%m-%d %H:%M:%S')
    timestamp2 = datetime.datetime.strptime(log_lines[i].split(".")[0], '%Y-%m-%d %H:%M:%S')

    # 计算时间间隔
    interval = timestamp2 - timestamp1

    # 将时间间隔添加到间隔时间列表中
    intervals.append(interval)

# 对间隔时间列表进行倒序排序
intervals.sort(reverse=True)

# 取前10条间隔时间
top_10_intervals = intervals[:10]
# 取前10条间隔时间的秒数
top_10_intervals_seconds = [int(interval.total_seconds()) for interval in top_10_intervals]

# 打印结果
print(top_10_intervals_seconds)

# benshanzhufu  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
# douyin_author [30, 28, 25, 20, 18, 18, 17, 17, 17, 16]
# douyin_recommend  [62, 50, 38, 34, 33, 31, 31, 31, 31, 31]
# kuaishou_author  [31, 27, 23, 21, 21, 21, 19, 19, 17, 15]
# kuaishou_recommend  [27, 23, 23, 23, 23, 22, 22, 22, 21, 21]
# xigua_author  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
# xigua_search  [18, 14, 14, 13, 13, 12, 11, 11, 11, 10]
# xiaoniangao_author  [4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
# xiaoniangao_play  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
# xiaoniangao_hour  [61, 3, 2, 2, 2, 2, 2, 2, 1, 1]
# suisuiniannian  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
# gzh1  [26, 26, 26, 25, 25, 25, 25, 24, 24, 24]
# gzh2  [28, 26, 25, 25, 24, 24, 24, 24, 24, 24]
# gzh3  [27, 26, 26, 25, 25, 25, 24, 24, 24, 24]
# gzh4  [26, 26, 25, 24, 24, 24, 24, 24, 24, 24]
# gzh5  [29, 26, 25, 25, 24, 24, 24, 24, 24, 24]
# gzh6  [26, 25, 25, 25, 25, 24, 24, 24, 24, 24]