Find unused helpers in a Rails app (slow)

find_unused_helpers.rb

#!/usr/bin/env ruby
#
# Shotgun approach (read: slow and dirty hack) to help find unused helpers in a Rails application
#

puts "Loading all source files into memory :("
source = {}

Dir["app/**/**/*.*"].each do |f|
  source[ f ] = File.readlines( f )
end

puts "#{source.size} files loaded into memory"


helpers = []

source.keys.grep(/app\/helpers/).each do |f|
  code = source[ f ]
  code.each do |line|
    if line =~ /def ([^\(\s]+)/
      helpers << $1.chomp
    end
  end
end

puts "Scanning for #{helpers.size} helpers"

# combine the source code
complete_code = source.values.flatten

# Find occurances
helpers.each do |name|
  found = false

  complete_code.each do |line|
    next if line =~ /def #{name}/
    found = true if line =~ /#{name}/
  end

  puts "No traces found of #{name}..." unless found
end

I wish this was updated for Ruby 1.9. I get an invalid byte sequence in UTF-8 (ArgumentError) error on your line =~ /dev #{name}/.

@moneypenny skipping asset files made me avoid it.

Thanks for this. My hack amendment to make it work in Ruby 2.0 was

  complete_code.each do |line|
    next if line != 'UTF-8'
    next if line =~ /def #{name}/
    found = true if line =~ /#{name}/
  end

I rewrote some of this to be much faster (about 80% faster in my case):

https://gist.github.com/kylefox/617b0bead5f53dc53a224a8651328c92

@cheshire137 As @markoa mentioned, you likely ran into the UTF errors while processing a binary file (i.e. a bitmap image). My update whitelists the file types to search, which avoids the UTF error and greatly speeds up the script overall.

@tylerlee FYI your change seems to cause every single helper to be flagged as unused because you're checking the line against the literal 'UTF-'8' string.