import threading # 多线程模块
import queue # 队列模块
import requests
from lxml import etree
import time
import random
import json
concurrent = 3 #采集线程数
conparse = 3 # 解析线程
class Parse(threading.Thread): # 解析线程类
# 初始化属性
def __init__(self,number,data_list,req_thread,f):
super(Parse ,self).__init__()
self.number = number # 线程编号
self.data_list = data_list # 数据队列
self.req_thread = req_thread # 请求队列,为了判断采集线程存活状态
self.f = f # 获取文件对象
self.is_parse = True # 判断是否从数据队列里提取数据
def run(self):
print('启动%d号解析线程' % self.number)
# 无限循环,
while True:
# 如何判断解析线程的结束条件
for t in self.req_thread: # 循环所有采集线程
if t.is_alive(): # 判断线程是否存活
break
else: # 如果循环完毕,没有执行break语句,则进入else
if self.data_list.qsize() == 0: # 判断数据队列是否为空
self.is_parse = False # 设置解析为False
# 判断是否继续解析
if self.is_parse: # 解析
try:
data = self.data_list.get(timeout=3) # 从数据队列里提取一个数据
except Exception as e: #超时以后进入异常
data = None
# 如果成功拿到数据,则调用解析方法
if data is not None:
self.parse(data) # 调用解析方法
else:
break # 结束while 无限循环
print('退出%d号解析线程' % self.number)
# 页面解析函数
def parse(self,data):
html = etree.HTML(data)
# 获取所有段子div
duanzi_div = html.xpath('//div[@id="content-left"]/div')
for duanzi in duanzi_div:
# 获取昵称
nick = duanzi.xpath('./div//h2/text()')[0]
nick = nick.replace('\n', '')
# 获取年龄
age = duanzi.xpath('.//div[@class="author clearfix"]/div/text()')
if len(age) > 0:
age = age[0]
else:
age = 0
# 获取性别
gender = duanzi.xpath('.//div[@class="author clearfix"]/div/@class')
if len(gender) > 0:
if 'women' in gender[0]:
gender = '女'
else:
gender = '男'
else:
gender = '中'
# 获取段子内容
content = duanzi.xpath('.//div[@class="content"]/span[1]/text()')[0].strip()
# 获取好笑数
good_num = duanzi.xpath('./div//span[@class="stats-vote"]/i/text()')[0]
# 获取评论
common_num = duanzi.xpath('./div//span[@class="stats-comments"]//i/text()')[0]
item = {
'nick': nick,
'age': age,
'gender': gender,
'content': content,
'good_num': good_num,
'common_num': common_num,
}
self.f.write(json.dumps(item,ensure_ascii=False) + '\n')
class Crawl(threading.Thread): # 采集线程类
# 初始化
def __init__(self,number,req_list,data_list):
# 调用Thread 父类方法
super(Crawl,self).__init__()
# 初始化子类属性
self.number = number
self.req_list = req_list
self.data_list = data_list
self.headers = {
'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.89 Safari/537.36'
}
# 线程启动的时候调用
def run(self):
# 输出启动线程信息
print('启动采集线程%d号' % self.number)
# 如果请求队列不为空,则无限循环,从请求队列里拿请求url
while self.req_list.qsize() > 0:
# 从请求队列里提取url
url = self.req_list.get()
print('%d号线程采集:%s' % (self.number,url))
# 防止请求频率过快,随机设置阻塞时间
time.sleep(random.randint(1,3))
# 发起http请求,获取响应内容,追加到数据队列里,等待解析
response = requests.get(url,headers=self.headers)
if response.status_code == 200:
self.data_list.put(response.text) # 向数据队列里追加 def main():
# 生成请求队列
req_list = queue.Queue()
# 生成数据队列 ,请求以后,响应内容放到数据队列里
data_list = queue.Queue()
# 创建文件对象
f = open('duanzi.json','w',encoding='utf-8')
# 循环生成多个请求url
for i in range(1,13 + 1):
base_url = 'https://www.qiushibaike.com/8hr/page/%d/' % i
# 加入请求队列
req_list.put(base_url)
# 生成N个采集线程
req_thread = []
for i in range(concurrent):
t = Crawl(i + 1,req_list,data_list) # 创造线程
t.start()
req_thread.append(t)
# 生成N个解析线程
parse_thread = []
for i in range(conparse):
t = Parse(i + 1,data_list,req_thread,f) # 创造解析线程
t.start()
parse_thread.append(t)
for t in req_thread:
t.join()
for t in parse_thread:
t.join()
# 关闭文件对象
f.close()
if __name__ == '__main__':
main()