From: Robert Klemme <shortcutter@...>
Date: 2007-05-26T17:25:03+09:00
Subject: Re: gsub + loop

On 26.05.2007 09:44, Eleo wrote:
> This question actually pertains to a Rails app but it's more of a
> general Ruby question so I'll ask it here.
> 
> Within a body of text I'm trying to match URLs for
> youtube/google/myspace/etc videos and replace them with their associated
> embed codes.  For each different site there is a different regular
> expression and a different embed code.  So I made a hash with each value
> being an array containing the regular expression and replacement like:
> {:videosite => [regexp, replacement]}.  From there I figured I should
> loop through the hash and check the text against each expression and
> replace the URLs if necessary.  Unfortunately I am having a problem:
> 
> http://pastie.caboo.se/64785
> 
> class Embedment < ActiveRecord::Base
>   def self.capture_embedments(text)
>     embedments = []
>     regexps.each do |k,v|
>       text.gsub!(v[0]) do |match|
>         embedments << embedment = self.new
>         embedment.html = v[1]
>       end
>     end
> 
>     return embedments
>   end
> 
>   def self.regexps
>     return {
>       :youtube    => [/\(youtube:.*?(?:v=)?([\w|-]{11}).*\)/, "<object
> width=\"425\" height=\"350\"><param name=\"movie\"
> value=\"http://www.youtube.com/v/#{match[0]}\"></param><param
> name=\"wmode\" value=\"transparent\"></param><embed
> src=\"http://www.youtube.com/v/#{match[0]}\"
> type=\"application/x-shockwave-flash\" wmode=\"transparent\"
> width=\"425\" height=\"350\"></embed></object>"]
>     }
>   end
> end
> 
> I think the flaw in my plan is that the second member of the array is
> being parsed as soon as it's referenced, so this raises an exception
> (undefined local variable or method `match' for Embedment:Class) since
> the object 'match' does not exist yet.

You're right on!

>  I'm guessing my approach to this
> problem is very very wrong, but I have yet to see past my poor solution.
> The reason why I separated the regular expressions from the method is
> because there's going to be a few tens of them and I wanted to
> consolidate them.  Any help would be appreciated.

You are not too far away.  Just use proper regexp escapes in the 
replacement string plus apply the replacement twice.   So #{match[0]} 
becomes \\& and #{match[1]} becomes \\1 etc.

class Embedment < ActiveRecord::Base
   def self.capture_embedments(text)
     embedments = []
     REGEXPS.each do |k,v|
       text.gsub!(v[0]) do |match|
         embedment = new
         embedments << embedment
         embedment.html = match.sub(*v)
       end
     end

     return embedments
   end
end


Note also, that it's better to make the replacements a constant in the 
class.  And if you iterate only, you don't need symbols as keys, just do

REGEXPS = {
   /\(youtube:.*?(?:v=)?([\w|-]{11}).*\)/ => "...",
}

And then

     REGEXPS.each do |rx,repl|
       text.gsub!(rx) do |match|
         embedment = new
         embedments << embedment
         embedment.html = match.sub(rx,repl)
       end
     end

If you need more flexibility you can replace the replacement string with 
a block.  Then you can do

class Embedment
   REGEXPS = {
     /\(youtube:.*?(?:v=)?([\w|-]{11}).*\)/ =>
       lambda {|match| "<!-- matched: #{match} -->" }
   }
end


Then you can do

     REGEXPS.each do |rx,repl|
       text.gsub!(rx) do |match|
         embedment = new
         embedments << embedment
         embedment.html = match.sub(rx,&repl)
       end
     end


Kind regards

	robert