From: Aaron Patterson <aaron@...>
Date: 2009-07-02T06:20:10+09:00
Subject: Re: converting the string

On Thu, Jul 02, 2009 at 05:29:04AM +0900, Vlad Smith wrote:
> Greg Willits wrote:
> > Vlad Smith wrote:
> >> Hi everybody! I`m sorry for asking a silly question .. i hope you`ll
> >> find some time to assist
> >> 
> >> while searching shrough html body i get the string :
> >> <td class=blk11 ><img
> >> src='https://sc.omniture.com/sc13_5/reports/chart.php?id=CPRIY_6NvJ_kj7s&s=www461.sj2&type=GIF'
> >> width=624 height=280 border=0 align=absmiddle
> >> usemap=#imIY_6NvJ_kj7s></td>
> >> 
> >> how can i modify it to get just a plain link without tags and params? :
> >> https://sc.omniture.com/sc13_5/reports/chart.php?id=CPRIY_6NvJ_kj7s&s=www461.sj2&type=GIF
> > 
> > You'll have to expand this to take care of all possible scenarios, but 
> > here's an example:
> > 
> >    x = "<img src=\"http://some_url_to_scrape\">"
> >    y = x.scan(/src=\"([\S\s]+?)\"/)
> > 
> > That will return an array, and you'll have to fish the string out of the 
> > array with y[0][0].
> > 
> > That works specifically with proper HTML using double quotes where your 
> > examples above used malformed single quotes, but you can either use 
> > multple expressions, or build a more complex one to cover the various 
> > cases of quotes href, src, and other attributes names, etc.
> > 
> > There's other ways to do it, this is just a small example to give you 
> > some ideas.
> > 
> > -- gw
> 
> Thanks! that worked!
> 
> i also accidently noticed a great feature taken from perl that worked 
> also:
> 
> x = <td class=blk11 ><img 
> src='https://sc.omniture.com/sc13_5/reports/chart.php?id=CPRIY_6NvJ_kj7s
> s=www461.sj2&type=GIF'width=624 height=280 border=0 
> align=absmiddleusemap=#imIY_6NvJ_kj7s></td>
> x = $1 if x =~ /.*(https.*GIF).*/

Please don't do this.  Every time you parse HTML with a regular
expression, a kitten dies.

Instead, try using an HTML parsing library:

x = <<-eohtml
  <td class=blk11 ><img 
  src='https://sc.omniture.com/sc13_5/reports/chart.php?id=CPRIY_6NvJ_kj7ss=www461.sj2&type=GIF'width=624 height=280 border=0 
  align=absmiddleusemap=#imIY_6NvJ_kj7s></td>
eohtml

puts Nokogiri::HTML(x).at('img')['src']

-- 
Aaron Patterson
http://tenderlovemaking.com/