From: William James <w_a_x_man@...>
Date: 2008-12-07T05:34:28+09:00
Subject: Re: Directory searching againist a text file

Stuart Clarke wrote:

> I am in the middle of writing a quick program which will scan the
> contents of a given file path recursively for a list of keywords
> stored in a file. My code so far is below, but before moving ahead I
> have two questions.
> 
> First: I am passing in a text file called "terms.txt" to search for
> each keyword in the file I assume the best way to to do so is as
> follows:
> 
> terms.each do |term|
> if line =~ term
>   puts ""
> end
> 
> My second question is: This program works well for searching text
> files but what about word docs and spreadsheets? Do i need some
> Windows API in there??

You can read these files if you open them in binary mode.
However, they will contain so much extra binary crap that
it may not be easy to search in them.

> 
> 
> Many thanks
> 
> 
> require 'find'
> 
> class ESearch
> 
>   #method which is passed file path from cmd line
>   def scanFiles(path)
>     terms  = "C:\Documents and Settings\user\Desktop\terms.txt"
>     #process each file under the passed file path
>     Find.find(path) do |curPath|
>       next unless File.file?(curPath)
>       #process the contens of each file line by line counting line
> nmbers
>       File.open(curPath) do |file|
>         file.each do |line|
>           #check if a line in the file matches term and output the
> path and line number
>           if line =~ terms
>             puts "#{curPath}"
>           end
>         end
>       end
>     end
>   end
> end
> 
> #run of cmd line pass in file path, this will ask for a file path if
> one is not passed
> if FILE == $0
>   if ARGV.size != 1
>     puts "Use: #{$0} [path]"
>     exit
>   end
> 
>   esearch = ESearch.new()
>   esearch.scanFiles(ARGV[0])
> end


terms = IO.read("terms.txt").strip.split(/\s*\n\s*/)

ARGF.each{|line| line.strip!
  if terms.include? line
    puts "#{ARGF.filename}:#{ARGF.lineno}: #{line}"
  end
}

Running it:

ruby scanner.rb *.dat