重写了connect通信程序与服务器通信系统,彻底重写了终端节点集群,对整个系统进行了较大幅度的优化

master
wufayuan 2 years ago
parent ae894c1fc0
commit f7c0dd043d

@ -1,13 +1,14 @@
import json import json
import multiprocessing
import socket import socket
import struct import struct
import sys import sys
import threading
from configparser import ConfigParser from configparser import ConfigParser
from json import JSONDecoder from json import JSONDecoder
from msedge.selenium_tools import Edge from msedge.selenium_tools import Edge
from msedge.selenium_tools import EdgeOptions from msedge.selenium_tools import EdgeOptions
sys.path.append(r'F:\Users\28587\dcs') sys.path.append(r'F:\Users\28587\dcs')
from dcs.tests.zhiwang import * from dcs.tests.zhiwang import *
from dcs.tools import message_process as mp from dcs.tools import message_process as mp
@ -76,7 +77,7 @@ def crawl(request_map) -> dict:
return result_map return result_map
class Client(threading.Thread): class Client(multiprocessing.Process):
def __init__(self, server_ip, server_port, local_ip, local_port): def __init__(self, server_ip, server_port, local_ip, local_port):
super(Client, self).__init__() super(Client, self).__init__()
self.server_ip = server_ip self.server_ip = server_ip

File diff suppressed because one or more lines are too long

@ -92,8 +92,8 @@ class Crawler(threading.Thread):
def run(self) -> None: def run(self) -> None:
try: try:
# self.crawl_zhiwang(user_name=self.partial_task.cui.user_name) self.crawl_zhiwang(user_name=self.partial_task.cui.user_name)
self.test_simulation(user_name=self.partial_task.cui.user_name) # self.test_simulation(user_name=self.partial_task.cui.user_name)
except Exception as e: except Exception as e:
logger.error(f'[ERROR] {str(e)}') logger.error(f'[ERROR] {str(e)}')
finally: finally:
@ -145,7 +145,6 @@ class Spider_task(threading.Thread):
pages_start = pages_end pages_start = pages_end
tmp['pages_end'] = pages_start tmp['pages_end'] = pages_start
self.partial_tasks.append(Spider_partial_task(self, tmp)) self.partial_tasks.append(Spider_partial_task(self, tmp))
logger.info(f'[TASK] all tasks: {self.partial_tasks}')
def is_all_task_crawled(self): def is_all_task_crawled(self):
for task in self.partial_tasks: for task in self.partial_tasks:

@ -137,7 +137,7 @@ function execute(cmd) { //调用cmd命令
}) })
} }
app.post('/check', function (req, res) { app.post('/check', function (req, res) {
execute('python connect.py --ip 127.0.0.1 --port 7777 crawling --word computer --cookie god --pages_start 1 --pages_end 3'); execute('python connect.py --ip 127.0.0.1 --port 7777 crawling --word computer --cookie god --pages_start 1 --pages_end 5');
fs.readFile('./result.json', 'utf-8', function (err, data) { fs.readFile('./result.json', 'utf-8', function (err, data) {
if (err) { if (err) {
console.error(err); console.error(err);

File diff suppressed because it is too large Load Diff
Loading…
Cancel
Save