以下是改进后的代码import requests
from selenium import webdriver
from pyquery import PyQuery
import re
import base64
word = input('输入爬取的单词图片:')
driver = webdriver.Edge()
url = f'https://www.google.com/search?q={word}&source=lnms&tbm=isch'
driver.get(url)
# input(":")
img = driver.page_source
length = 0
for i in PyQuery(img)('#islrg div div a div img').items():
img_name = '%s.jpg' % word
if not i.attr('src'):
code = requests.get(i.attr('data-src')).content
else:
if str(i.attr('src')).startswith('http'):
code = requests.get(i.attr('src')).content
else:
code = base64.b64decode(re.sub(r'data:image/.+?;base64,', '', str(i.attr('src'))))
with open(img_name, 'wb') as f:
f.write(code)
break
非常感谢,大大提升了搜索效率 |