Python, BeautifulSoup or LXML - Parsing image URL's from HTML using CSS tags

Question

I have searched high and low for a decent explanation of how BeautifulSoup or LXML work. Granted, their documentation is great, but for someone like myself, a python/programming novice, it is difficult to decipher what I am looking for.

Anyways, as my first project, I am using Python to parse an RSS feed for post links - I have accomplished this with Feedparser. My plan is to then scrape each posts' images. For the life of me though, I can not figure out how to get either BeautifulSoup or LXML to do what I want! I have spent hours reading through the documentation and googling to no avail, so I am here. The following is a line from the Big Picture (my scrapee).

<div class="bpBoth"><a name="photo2"></a><img src="http://inapcache.boston.com/universal/site_graphics/blogs/bigpicture/shanghaifire_11_22/s02_25947507.jpg" class="bpImage" style="height:1393px;width:990px" /><br/><div onclick="this.style.display='none'" class="noimghide" style="margin-top:-1393px;height:1393px;width:990px"></div><div class="bpCaption"><div class="photoNum"><a href="#photo2">2</a></div>In this photo released by China's Xinhua news agency, spectators watch an apartment building on fire in the downtown area of Shanghai on Monday Nov. 15, 2010. (AP Photo/Xinhua) <a href="#photo2">#</a><div class="cf"></div></div></div>

So, according to my understanding of the documentation, I should be able to pass the following:

soup.find("a", { "class" : "bpImage" })

To find all instances with that css class. Well, it doesn't return anything. I'm sure I'm overlooking something trivial so I greatly appreciate your patience.

Thank you very much for your responses!

For future googlers, I'll include my feedparser code:

#! /usr/bin/python

# RSS Feed Parser for the Big Picture Blog

# Import applicable libraries

import feedparser

#Import Feed for Parsing
d = feedparser.parse("http://feeds.boston.com/boston/bigpicture/index")

# Print feed name
print d['feed']['title']

# Determine number of posts and set range maximum
posts = len(d['entries'])

# Collect Post URLs
pointer = 0
while pointer < posts:
    e = d.entries[pointer]
    print e.link
    pointer = pointer + 1

Answer 1

Using lxml, you might do something like this:

import feedparser
import lxml.html as lh
import urllib2

#Import Feed for Parsing
d = feedparser.parse("http://feeds.boston.com/boston/bigpicture/index")

# Print feed name
print d['feed']['title']

# Determine number of posts and set range maximum
posts = len(d['entries'])

# Collect Post URLs
for post in d['entries']:
    link=post['link']
    print('Parsing {0}'.format(link))
    doc=lh.parse(urllib2.urlopen(link))
    imgs=doc.xpath('//img[@class="bpImage"]')
    for img in imgs:
        print(img.attrib['src'])

Answer 2

The code you have posted looks for all a elements with the bpImage class. But your example has the bpImage class on the img element, not the a . You just need to do:

soup.find("img", { "class" : "bpImage" })

Answer 3

Using pyparsing to search for tags is fairly intuitive:

from pyparsing import makeHTMLTags, withAttribute

imgTag,notused = makeHTMLTags('img')

# only retrieve <img> tags with class='bpImage'
imgTag.setParseAction(withAttribute(**{'class':'bpImage'}))

for img in imgTag.searchString(html):
    print img.src