爬虫大作业

# -*- coding: utf-8 -*-
import requests
from bs4 import BeautifulSoup as bs

 

def gettext(url):
header = {
'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.117 Safari/537.36'}
html = requests.get(url, headers=header).content

soup = bs(html, 'html.parser')
info = soup.select('div.newList.black01 a')
a = []
for i in info:
a.append(i.text)
print(i.text)
return a


if __name__ == '__main__':
url = "http://www.sontan.net/newsCenter.do"
#html = getreq(url)
info = gettext(url)
print(info)
for i in info:
print(i)
f = open('i.txt', 'a+',encoding='utf-8')
f.write(i)
f.write('\n')
f.close()

 

import jieba
import PIL
from wordcloud import WordCloud
import matplotlib.pyplot as p
import os

info = open('i.txt', 'r', encoding='utf-8').read()
text = ''
text += ' '.join(jieba.lcut(info))
wc = WordCloud(font_path='C:\Windows\Fonts\STZHONGS.TTF', background_color='White', max_words=50)
wc.generate_from_text(text)
p.imshow(wc)
# p.imshow(wc.recolor(color_func=00ff00))
p.axis("off")
p.show()
wc.to_file('dream.jpg')

 

# -*- coding: utf-8 -*-
import requests
from bs4 import BeautifulSoup as bs

 

def gettext(url):
header = {
'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.117 Safari/537.36'}
html = requests.get(url, headers=header).content

soup = bs(html, 'html.parser')
info = soup.select('div.newList.black01 a')
a = []
for i in info:
a.append(i.text)
print(i.text)
return a


if __name__ == '__main__':
url = "http://www.sontan.net/newsCenter.do"
#html = getreq(url)
info = gettext(url)
print(info)
for i in info:
print(i)
f = open('i.txt', 'a+',encoding='utf-8')
f.write(i)
f.write('\n')
f.close()

 

import jieba
import PIL
from wordcloud import WordCloud
import matplotlib.pyplot as p
import os

info = open('i.txt', 'r', encoding='utf-8').read()
text = ''
text += ' '.join(jieba.lcut(info))
wc = WordCloud(font_path='C:\Windows\Fonts\STZHONGS.TTF', background_color='White', max_words=50)
wc.generate_from_text(text)
p.imshow(wc)
# p.imshow(wc.recolor(color_func=00ff00))
p.axis("off")
p.show()
wc.to_file('dream.jpg')

 

 

posted @ 2018-04-26 15:04  177黄南龙  阅读(167)  评论(0编辑  收藏  举报