def __init__(self):
self.dburl = MongoUrl()
self.dbarticle = MongoArticle()
self.url_set = set()
self.headers = {
"User-Agent":
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.12 Safari/537.36"
}
def __init__(self, master):
self.window = master
sw = self.window.winfo_screenwidth()
sh = self.window.winfo_screenheight()
ww = 1400
wh = 650
x = (sw - ww) / 2
y = (sh - wh) / 2
self.window.geometry('%dx%d+%d+%d' % (ww, wh, x, y)) # 父容器大小
self.threadnumVar = tk.IntVar()
self.timeVar = tk.IntVar()
self.save_pathVar = tk.StringVar()
self.logMessage = JoinableQueue()
self.errMessage = JoinableQueue()
self.dbconf = MongoConfig()
self.dburl = MongoUrl()
self.dbarticle = MongoArticle()
self.create_page()
self.show_logs()
self.asyCraler()
#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
__author__ = 'AJay'
__mtime__ = '2019/5/9 0009'
"""
from db import MongoArticle, MongoUrl
mu = MongoUrl()
ma = MongoArticle()
import time
from shortuuid import uuid
import os
#TODO:按照随机的文件名，导出每天的新闻到对应当天时间的文件夹中
'''
输入一个路径、如果路径存在、则使用路径、如果路径不存在则使用文件的路径。输出文件路径所在的位置
导出当天的新闻
'''
class EexportTxt():
def __init__(self):
self.base_path = os.path.abspath(os.path.dirname(__file__))
self.ds = 0
self.length_p = 30
self.file_size = 300 * 1024 # 300k
def _is_input_path(self, input_path):
if not os.path.exists(input_path): # 路径函数
def __init__(self):
self.dburl = MongoUrl()
self.dbarticle = MongoArticle()
self.url_set = set()
self.url_queue = queue.Queue()
self.init_set()