需要实现的功能：给出一个网站列表，抓出这些网页上的图片。
实现方式： 下载网页源码，在源码中识别包含图片url的标签，如<img>,<div>,<li>。由于对html了解较少，哪些标签可能含有图片是从查看多个网站的源码中总结出来的。
调用的库：Selenium（加载Chrome驱动）--获取执行JS后的源码。
threading--实现多进程
代码：
from urllib.parse import urljoin,urlparse
import os
import threading
from time import ctime
from selenium import webdriver
import re
class myThread(threading.Thread):
def __init__(self,func,args,name=‘‘):
threading.Thread.__init__(self)
self.name=name
self.func=func
self.args=args
self.is_end=False
def getResult(self):
return self.res
def run(self):
self.res=self.func(*self.args)
def filter_in_tag(page_file,tag):
url_in_tag = []
url_in_tag.append(‘------------------%s--------------------‘ % (tag))
with open(page_file, ‘r‘, encoding=‘utf-8‘) as jj:
for line in jj:
##先找出li所有标签
reg = ‘<%s [^>]*>‘ % (tag)
all_tag_str = re.findall(reg, line)
for tag_str in all_tag_str:
if re.search(‘https?://[^\‘\"\)]+‘, tag_str):
url_in_tag.extend(re.findall(‘http?://[^\‘\"]+‘, tag_str))
return url_in_tag
def process(m_url):
imgs,big_files,hrefs=[],[],[]
##先找出图片
##添加参数，使chrome不出现界面
chrome_options = webdriver.chrome.options.Options()
chrome_options.add_argument(‘--headless‘)
chrome_options.add_argument(‘--disable-gpu‘)
driver = webdriver.Chrome(r‘C:\Program Files (x86)\Google\Chrome\Application\chromedriver.exe‘,
chrome_options=chrome_options) ##driver = webdriver.PhantomJS(executable_path=‘/bin/phantomjs/bin/phantomjs‘)#如果不方便配置环境变量。就使用phantomjs的绝对路径也可以
driver.set_page_load_timeout(30)
try:
driver.get(m_url)
except Exception as e: ##(HTTPError,URLError,UnicodeDecodeError,WindowsError) as e:
err_info=‘url open error: %s\n, reason: %s\n‘%(m_url,e)
print(err_info)
err_log.write(err_info)
#print(‘url open error: %s\n, reason: %s‘%(m_url,e))
return []
imgs = []
imgs.append(‘------------------<img src=>-----------------‘)
for x in driver.find_elements_by_tag_name("img"):
imgs.append(x.get_attribute(‘src‘))
# 找出所有div li标签中的链接
with open(‘tmp_page_source.html‘,‘w‘,encoding=‘utf-8‘) as tmp_f:
tmp_f.write(driver.page_source)
for tag in (‘li‘, ‘div‘):
imgs.extend(filter_in_tag(‘tmp_page_source.html‘,tag))
##列表去重复
imgs_uniq = []
for url in imgs:
if (url not in imgs_uniq) and (url): ##url不在新列表中且url不为空
imgs_uniq.append(url)
##查找页面中的a链接中的大文件和其它网页
links=[a_link.get_attribute(‘href‘) for a_link in driver.find_elements_by_tag_name(‘a‘) if a_link.get_attribute(‘href‘)]
driver.quit()
for link in links:
host = urlparse(m_url).netloc.split(‘@‘)[-1].split(‘:‘)[0]
dom = ‘.‘.join(host.split(‘.‘)[-2:])
if link.startswith(‘mailto:‘):
continue
if not link.startswith(‘http‘):
link=urljoin(m_url,link)
f_name = urlparse(link).path.split(‘/‘)[-1]
f_type = os.path.splitext(f_name)[1]
if f_type not in (‘.htm‘,‘.html‘,‘shtml‘,‘‘):
big_files.append(link)
continue
if link in seen_links:
pass#print(link,‘--aleady processed,pass.‘)
else:
if dom not in link:
pass#print(link,‘--not in domain,pass.‘)
else:
hrefs.append(link)
seen_links.append(link)
return imgs_uniq,big_files,hrefs
##对process处理结果进行分析，得出如下统计数据：
##图片：100，HTTP协议占比：80%，HTTP协议下各种后缀的数量：jpg-50，gif-30
##大文件：10，HTTP协议占比：100%，HTTP协议下各种后缀的数量：pdf-10
def ret_analyse(url_list):
to_len=len(url_list)##含有3行标识信息，非url
http_list= [url for url in url_list if url.startswith("http://")]
http_perc=‘%.1f%%‘%(len(http_list)/to_len*100) if to_len>0 else ‘0‘
exts_dict={}
for url in url_list:
if url.startswith(‘-----------‘): ##排除‘-------img:src-----’等
continue
f_name = urlparse(url).path.split(‘/‘)[-1]
f_type = os.path.splitext(f_name)[1]
if f_type not in exts_dict:
exts_dict[f_type]=1
else:
exts_dict[f_type]+=1
return to_len,http_perc,exts_dict
##对一组url调用process函数处理，并输出结果到文本
def group_proc(url_f , urls,is_analyse) :
links=[] ##存储该页面除大文件外的a链接
##定义写日志的函数
def wLog(*lines):
for line in lines:
try:
url_f.write(line + ‘\n‘)
except Exception as e:
print(‘write eror,line:%s, err: %s‘%(line,e))
for url in urls:
proc_ret=process(url)
if proc_ret:
img_list,bigfile_list,link_list=proc_ret
wLog(‘*‘*40,‘from: ‘,url) # 分隔行+起始行
if is_analyse:
img_output=‘图片：%d，HTTP协议占比：%s，HTTP协议下各种后缀的数量：%s‘%(ret_analyse(img_list)[0]-3,ret_analyse(img_list)[1],ret_analyse(img_list)[2]) ##图片含有3行标识信息
big_output = ‘大文件：%d，HTTP协议占比：%s，HTTP协议下各种后缀的数量：%s‘ % (ret_analyse(bigfile_list))
wLog(img_output,big_output)
img_list = ‘\n‘.join(img_list)
bigfile_list = ‘\n‘.join(bigfile_list)
wLog(‘imgs:‘,img_list,‘bigfiles: ‘,bigfile_list,‘*‘*40)
imgs_f.write(img_list + ‘\n‘)
if bigfile_list:
bigfiles_f.write(bigfile_list + ‘\n‘)
if link_list:
links.extend(link_list)
return links
def main(depth):
u_file=open(‘urls.txt‘,‘r‘)
links=[line.strip(‘\n‘) for line in u_file]
links=[‘http://‘+link for link in links if not link.startswith(‘http‘)]
u_file.close()
for i in range(depth):
is_analyse=True if i==0 else False ##对第一层数据需要分析统计
url_f = open(‘layer‘ + str(i)+‘.txt‘,‘w‘)
next_links=[]
if not links:
break
else:
print(‘第 %d 层开始爬取...‘%(i))
##将链接分配给5组
avg=len(links)//5
links_grp=[]
if avg==0:
grp_len=len(links)
for i in range(grp_len):
links_grp.append([links[i]])
else:
grp_len = 5
links_grp=links[:avg],links[avg:avg*2],links[avg*2:avg*3],links[avg*3:avg*4],links[avg*4:]
#for i in range(grp_len):
#url_f.write(‘link_group %d:%s‘%(i,links_grp[i]))
##新建5个线程，分别处理5组url
threads=[]
for i in range(grp_len):
t=myThread(group_proc,(url_f,links_grp[i],is_analyse),group_proc.__name__)
threads.append(t)
##线程同时启动
for i in range(grp_len):
print(‘线程%d开始运行,时间：%s‘%(i,ctime()))
threads[i].setDaemon(True)
threads[i].start()
##等待线程结束，结束后将各组url中获取的外链加入到下一次处理的列表中
for i in range(grp_len):
threads[i].join()
print(‘线程%d运行结束,时间：%s‘ % (i, ctime()))
ret_links=threads[i].getResult()
next_links.extend(ret_links)
links=next_links
url_f.close()
if __name__==‘__main__‘:
seen_links = []
imgs_f = open(‘图片.txt‘, ‘w‘,encoding=‘utf-8‘)
bigfiles_f = open(‘大文件.txt‘, ‘w‘,encoding=‘utf-8‘)
err_log = open(‘err_log.txt‘, ‘w‘,encoding=‘utf-8‘)
depth=int(input(‘请输入爬取深度：‘))
main(depth)
err_log.close()
imgs_f.close()
bigfiles_f.close()
input(‘按任意键退出...‘)
原文地址：https://www.cnblogs.com/qinfei88/p/10113951.html
时间： 2024-10-12 01:40:36