Server-side jquery

Question

Say I have a script written in perl or python. What's the easiest way to write a function that would use jquery selectors on strings as part of it? ie to be able to do:

jquery_selector('table.new#element', text)

where jquery_selector is a function that runs a jquery selector on the html string stored in text . Even if it was just limited to returning strings (not full jquery objects), it would still be really useful. ie if you were required to give a javascript function as a callback which would render the results to something comprehensible in the scripting language:

jquery_selector('table.new#element, text, 'function(e){return e.val()}')

And it would return the results of the callback as a list.

I realize that there are dom libraries for most languages, but jquery is so much better than most of them.

I am not asking about native libraries which have a syntax like jquery. I guess what would be needed is an API to a browser which jquery would run on? Or is this what node.js does?

Answer 1

if i understand your question correctly you want is something like phantom.js . PhantomJS is a headless WebKit with JavaScript API. you can inject jquery into it and use all the jquery selectors to manipulate the dom. you can make it work like a standalone server aswell.

Answer 2

I am confused as you have tagged the question with "node.js" though you refer to Python or Perl in your question. Running node.js and/or phantom.js just to run a selector on an HTML DOM sounds quite heavyweight to me, and as always introducing whole chains of dependencies should be considered carefully in real-world projects.

So for Python I would suggest running a combination of BeautifulSoup and soupselect , as mentioned in this answer . You can then do things like:

from BeautifulSoup import BeautifulSoup as Soup
from soupselect import select
import urllib

soup = Soup(urllib.urlopen('http://slashdot.org/'))
select(soup, 'div.title h3')

Note that soupselect seems to implement only a subset of jquery's CSS3 selectors, so for things like eg sibling selectors or pseudo-classes it may not work. In this case I would advise to consider porting the relevant part of the project to node.js were you can run Sizzle (jQuery's selector engine) or cheerio standalone in a somewhat lightweight environment.