From: Gregory Seidman <gsslist+ruby@...>
Date: 2006-10-12T02:49:33+09:00
Subject: Re: Yet another Hpricot question

On Thu, Oct 12, 2006 at 02:28:13AM +0900, Aaron Patterson wrote:
} On Thu, Oct 12, 2006 at 02:13:07AM +0900, Rick DeNatale wrote:
} > I'm trying to scan an html file using Hpricot to produce a table of
} > links within the file.
} > 
} > Right now I've got something like this.
} > 
} >   doc = Hpricot(open(url)).
} >   doc.search('a').each do | element |
} >         puts "#{element.inner_html}
} >         puts "     #{element.attributes['href']
} >   end
} > 
} > This works, but in this document some of the a tags use markup on
} > their contents. Something like
} > <a href="http://blah.org/blah.htm"><b>blah blah</b> blah</a>
} > 
} > I'd like to strip out the markup tags so that I'd get
} > 
} > blah blah blah
} >    http://blah.org/blah.htm
} > 
} > Is there some way to search for or iterate over the leaf elements of
} > the tree rooted by an element in Hpricot?
} 
} I had to do something similar in Mechanize, and this is what I came up
} with:
} 
}   class Hpricot::Elem
}     def all_text
}       text = ''
}       children.each do |child|
}         if child.respond_to? :content
}           text << child.content
}         end
}         if child.respond_to? :all_text
}           text << child.all_text
}         end
}       end
}       text
}     end
}   end
}   
}   doc = Hpricot("<a href=\"http://blah.org/blah.htm\"><b>blah blah</b> blah</a>")
}   doc.search('a').each do |e|
}     puts "#{e.all_text}"
}     puts "     #{e.attributes['href']}"
}   end

There is a simpler implementation of all_text:

class Hpricot::Elem
  def all_text
    text = ''
    traverse_text {|t| text << t.content }
    text
  end
end

} Hope that helps!
} --Aaron
--Greg