Extracting a div from parsed HTML

Question

It seems that lxml/etree are generally imported as from lxml import etree -- why is that? It keeps the code tidier, and while the potential namespace ambiguity might not be a concern, I don't have any incentive of doing this as it's generally frowned upon.

I know for a script of this size it doesn't matter much, but I'm going to be using these modules for a lot more. I'm also curious about what others have to say.

#!/usr/bin/python
# Stuart Powers http://sente.cc/
import sys
import urllib
import lxml.html
from cStringIO import StringIO
""" This script parses HTML and extracts the div with an id of 'search-results':
 ex: <div id='search-results'>...</div>
$ python script.py "http://www.youtube.com/result?search_query=python+stackoverflow&page=1"
The output, if piped to a file would look like: http://c.sente.cc/E4xR/lxml_results.html
"""
parser = lxml.html.HTMLParser()
filecontents = urllib.urlopen(sys.argv[1]).read()
tree = lxml.etree.parse(StringIO(filecontents), parser)
node = tree.xpath("//div[@id='search-results']")[0]
print lxml.etree.tostring(tree, pretty_print=True)

Answer 1

You might be confusing from lxml import etree that is a legitimate (even preferred) form of an absolute import with relative imports for intra-package imports that are discouraged: http://www.python.org/dev/peps/pep-0008/ (see "Imports" section)

Answer 2

In your and most of the cases I had while working with lxml.etree or lxml.html, there was only need for parsing and dumping, which in case of string input and output can be achieved with fromstring() and tostring() functions:

from lxml.html import fromstring, tostring

Which would transform your code to:

import sys
import urllib
from lxml.html import fromstring, tostring
data = urllib.urlopen(sys.argv[1]).read()
tree = fromstring(data)
node = tree.xpath("//div[@id='search-results']")[0]
print(tostring(tree, pretty_print=True))