From: Peter Szinek <peter@...>
Date: 2006-12-13T22:58:34+09:00
Subject: Re: Hpricot html parsing

Dhanasekaran Vivekanandhan wrote:
> yes, I want the text of the first <p> because it
> has an image. and reject if <p> has no image.
> thanks,
I see. Try this:
===============================================
require 'rubygems'
require 'hpricot'

doc = Hpricot %q{<p class=posted>
 this is fun
 <img src="" class="dhans"/>
</p>
<p class=posted>
 NO FUN
</p>
<p class=posted>
  fun again!
  <img src=""/>
</p>
<p class=posted>
  NO FUN AT ALL!
</p>
}

paragraphs = doc/'p'

good_elems = paragraphs.map.reject {|elem|  ((elem/"img").empty?) }
good_elems.each { |elem| puts elem.inner_text.strip }
===============================================

output:

************
this is fun
fun again!
************


You will need hpricot 0.4.84 because of inner_text - if you don't want
to install it (I did not experience any difficulties, so I can recommend
it) then you have to roll your own inner_text, but I guess this is not a
big problem.

Cheers,
Peter