From: Esmail Bonakdarian <ebonak-a@...>
Date: 2006-09-30T11:45:05+09:00
Subject: Re: newbie: how to find & extract a string from a file

Hi,

Thank you so much for these pointers. Am I correct in assuming
that REXML comes as part of standard Ruby? If so I think I will
go that route first.

I could cobble something together using various Linux tools
(grep and awk come to mind), but I want something in Ruby
(because I want to learn it) and also because it will be
more portable, for instance to the XP platform.

I appreciate you taking the time to post this and the references.
If you have any other ideas/approaches, I'm game.

Thanks again,

Esmail

MonkeeSage wrote:
> Esmail Bonakdarian wrote:
>> I would like to extract the filename with the .jpg extension.
>> What is the best approach for this?
> 
> Hi there,
> 
> You could use raw regexps and do it yourself, but you should probably
> use an HTML parser to extract HTML data. ;)
> 
> A nice HTML parser is Hpricot [1], but it requires an extension (you
> cen get it very easily via gems, see the link below). It is very easy
> to use, and it's fast.
> 
> Using Hpricot, you can do something like this:
> 
> require 'hpricot'
> require 'open-uri'
> soc = open('http://utopia.utexas.edu/maps/ireland.html')
> doc = Hpricot(soc)
> soc.close
> doc.search('//a').each { |elem|
>   href = elem.attributes['href']
>   if not href.nil? and
>      ['.jpg', '.jpeg'].include?(File.extname(href))
>     puts href
>   end
> }
> 
> Note that you can also use the built-in REXML parser [2], and do
> something like:
> 
> require 'rexml/document'
> require 'open-uri'
> include REXML
> soc = open('http://utopia.utexas.edu/maps/ireland.html')
> doc = Document.new(soc)
> soc.close
> doc.elements.each('//a') { |elem|
>   href = elem.attributes['href']
>   if not href.nil? and
>      ['.jpg', '.jpeg'].include?(File.extname(href))
>     puts href
>   end
> }
> 
> [1] http://code.whytheluckystiff.net/hpricot/
> [2] http://www.germane-software.com/software/rexml/docs/tutorial.html
> 
> Regards,
> Jordan
>