I try | MarsW

# encoding: utf-8

import urllib2
from lxml import etree

request = urllib2.Request("http://blog.marsw.tw")
response = urllib2.urlopen(request)
html = response.read()

page = etree.HTML(html)
count = 0
for url in page.xpath(u"//img/@src"):
    if url.startswith("http:"):
        print url
        
        try:
            img_request = urllib2.Request(url)
            img_response = urllib2.urlopen(img_request)
            img = img_response.read()
        except:
            continue

        count = count+1
        filename = str(count)+".jpg"
        pic_out = file(filename,'w')
        pic_out.write(img)
        pic_out.close()