用类方法写爬虫-爬东野圭吾白夜行小说第一章练手

import requests
import lxml.html
from multiprocessing.dummy import Pool
class Novespider(object):
    def __init__(self):
        self.novel_dict={}
    def get_content(self,url):
        html = requests.get(url).content
        self.get_novel(html)
    def get_novel(self,html):
        secetor = lxml.html.fromstring(html)
        data = secetor.xpath('//div[@class="readtext"]')
        for each_paragraph in data:
            title = each_paragraph.xpath('h4/text()')[0]
            # print(title)
            paragraph = each_paragraph.xpath('p/text()')
            #print(paragraph)
            paragraph = '\n'.join(paragraph)
        self.novel_dict[title]=paragraph
        #novel[title]=paragraph
    def write_txt(self):
        novel_paragraph_list=sorted(self.novel_dict.items(),key=lambda x:x[0])
        with open('东野圭吾.txt','w',encoding='utf-8') as file:
            for i in novel_paragraph_list:
                file.write(i[0])
                file.write('\n')
                file.write(i[1])
                file.write('\n')
if __name__=='__main__':
    urls = ['http://www.dongyeguiwu.com/books/baiyexing/53.html/{}'.format(str(i)) for i in  range(1,8,1)]
    #for i in url:
        #print(i)
    pool = Pool()
    data = Novespider()
    pool.map(data.get_content, urls)
    data.write_txt()
    print('搞掂了')

效果图片

用类方法写爬虫-爬东野圭吾白夜行小说第一章练手_第1张图片

你可能感兴趣的:(用类方法写爬虫-爬东野圭吾白夜行小说第一章练手)