crf 0.0.7 → 0.0.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: cedcfa93b22235198b5fe6cb7390bdee22f02109
-  data.tar.gz: fe5131cf2039948c9ffa7db13a7ced10d8f6f1a2
+  metadata.gz: 3b6b40091e9563035ada765e0622007dd18d27fc
+  data.tar.gz: 752bac46aef9db45980de0c18bbb65d8ac93cd16
 SHA512:
-  metadata.gz: 48487f353a617433b44ffc1514dc92fed3e365e4422a626e8ba28bbffb82b5859088649599e7fb142198d96174ecb7ffd95b41daf93766b8def8a36d2852ac11
-  data.tar.gz: f1a8b5badf11ede2776637900365fb49b132f34fcd65c02da7a2ed6b0772ed01d2484eca4f0a147d1951db4be8366b3b1644d7231545101054469e332f4a69d5
+  metadata.gz: 7111b4a41c5a117b54ea62f2e4a73444c680b26c9630c1b97425e09ff6974a6203150510a6abdc8def73a8fe90ca57520b0fe1e6dc62ea2abd5011895ac6bf9b
+  data.tar.gz: 1740bd54e2394b4e72bb04b25327a4a15264ca1eb4094e278e69d02883495b78f7cbb7ffd2bb44d9e7bf6bceead806c012896bd290b043655b977f95b4040674

data/.rspec

@@ -0,0 +1 @@
+--color

data/.rubocop.yml

@@ -4,7 +4,7 @@ AllCops:
     - bin/crf
 
 Documentation:
-  Enabled: true
+  Enabled: false
 
 LineLength:
   Max: 99

data/.travis.yml

@@ -1,6 +1,5 @@
 language: ruby
 rvm:
-  - 1.9.3
   - 2.0.0
   - 2.1
   - 2.2

data/Gemfile

@@ -3,7 +3,7 @@ source 'https://rubygems.org'
 # Specify your gem's dependencies in crf.gemspec
 gemspec
 
-gem 'rake', '~> 10.0'
+gem 'rake', '~> 11.1'
 gem 'rspec', '~> 3.4'
 
 group :test do

data/bin/crf

@@ -1,6 +1,7 @@
 #!/usr/bin/env ruby
 
 require 'crf'
+require 'crf/version'
 require 'optparse'
 
 options = { interactive: true, progress: true, fast: false }
@@ -20,6 +21,11 @@ parser = OptionParser.new do |opts|
     options[:progress] = false
   end
 
+  opts.on('-v', '--version', 'Displays version') do
+    puts "CRF Version: #{Crf::VERSION}"
+    exit
+  end
+
   opts.on('-h', '--help', 'Displays help') do
     puts opts
     exit
@@ -28,8 +34,11 @@ end
 
 parser.parse!
 
-if !ARGV[0].nil? && File.directory?(ARGV[0])
-  Crf::Checker.new(ARGV[0], options).check_repeated_files
+paths = []
+ARGV.each { |arg| paths << arg if File.directory?(arg) }
+
+unless paths.empty?
+  Crf::Checker.new(paths, options).check_repeated_files
 else
   STDOUT.puts 'No directory specified.'
 end

data/crf.gemspec

@@ -23,6 +23,6 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'ruby-progressbar', '~> 1.7', '>= 1.7.0'
 
   spec.add_development_dependency 'bundler', '>= 1.3.0', '< 2.0'
-  spec.add_development_dependency 'byebug' if RUBY_VERSION >= '2.0.0'
-  spec.add_development_dependency 'rubocop'
+  spec.add_development_dependency 'byebug', '~> 9.0', '>= 9.0.5' if RUBY_VERSION >= '2.0.0'
+  spec.add_development_dependency 'rubocop', '~> 0.38', '>= 0.37.2'
 end

data/lib/crf.rb

@@ -1,3 +1,4 @@
+require 'crf/version'
 require 'crf/finder'
 require 'crf/interactive_finder'
 require 'crf/remover'
@@ -7,30 +8,22 @@ require 'crf/configuration'
 require 'colorize'
 
 module Crf
-  #
-  # This class is the Crf starting point.
-  #
   class Checker
-    #
-    # The path where it will look for repetitions, the options provided, the repetitions found
-    # and the logger files are accesible from the outside and used in the class.
-    #
-    attr_reader :path, :options, :repetitions, :logger
+    attr_reader :paths, :options, :repetitions, :logger
 
     #
-    # Creates the object saving the directory's path and options provided. Options are set to
+    # Creates the object saving the paths and options provided. Options are set to
     # default if they are not given. It also creates the logger file.
     #
-    def initialize(path, options = { interactive: false, progress: false, fast: false })
-      @path = path
+    # @param path [Arsray] array of paths where the scan will start
+    # @param options [Hash] hash indicating the options of the scan
+    #
+    def initialize(paths, options = { interactive: false, progress: false, fast: false })
+      @paths = paths
       @options = options
       @logger = Crf::Logger.new
     end
 
-    #
-    # Starting point of Crf. You should call this if you want to check if a directory has
-    # duplicated files inside.
-    #
     def check_repeated_files
       find_repetitions
       return no_repetitions_found if repetitions.empty?
@@ -40,15 +33,18 @@ module Crf
     private
 
     def find_repetitions
-      logger.write "Looking for repetitions in #{path}"
-      finder = if options[:progress]
-                 Crf::InteractiveFinder.new(path, options[:fast])
-               else
-                 Crf::Finder.new(path, options[:fast])
-               end
+      logger.write "Looking for repetitions in #{paths}"
       @repetitions = finder.search_repeated_files
     end
 
+    def finder
+      unless instance_variable_defined?(:@finder)
+        @finder = Crf::InteractiveFinder.new(paths, options[:fast]) if options[:progress]
+        @finder = Crf::Finder.new(paths, options[:fast]) unless options[:progress]
+      end
+      @finder
+    end
+
     def no_repetitions_found
       logger.write 'No repetitions found'
       STDOUT.puts 'No repetitions found'.blue
@@ -62,9 +58,8 @@ module Crf
     end
 
     def remove_repetitions
-      remover = Crf::Remover.new(repetitions, logger) unless options[:interactive]
-      remover = Crf::InteractiveRemover.new(repetitions, logger) if options[:interactive]
-      remover.remove
+      return Crf::InteractiveRemover.new(repetitions, logger).remove if options[:interactive]
+      Crf::Remover.new(repetitions, logger).remove
     end
 
     def number_to_human_size(size)

data/lib/crf/configuration.rb

@@ -1,6 +1,3 @@
-#
-# Crf configuration
-#
 module Crf
   GEM_NAME = 'CRF'.freeze
   LOGGER_DATE_TIME_FORMAT = '%Y-%m-%d %H:%M:%S'.freeze

data/lib/crf/finder.rb

@@ -1,30 +1,24 @@
 require 'crf/repetitions_list'
 require 'digest'
 require 'ruby-progressbar'
+require 'byebug'
 
 module Crf
-  #
-  # This class finds the paths of all the repeated files inside the path passed as argument.
-  # All files repeated have the same file_identifier and file_hash.
-  #
   class Finder
-    #
-    # The original path provided and the list of files inside it are accessible from the outside.
-    #
-    attr_reader :path, :paths, :repetitions
+    attr_reader :paths, :repetitions, :files
 
-    #
+    ##
     # Creates the Finder object with a directory where it will look for duplicate files.
     # Path is the string representation of the absolute path of the directory.
     #
-    def initialize(path, fast = false)
-      @path = path
+    # @param paths [Array] paths of the folders where the scan will start.
+    # @param fast [Boolean] boolean indicating if this class will make a fast scan or not.
+    #
+    def initialize(paths, fast = false)
+      @paths = paths
       @fast = fast
     end
 
-    #
-    # Method that looks for the repeated files in the path specified when the object was created.
-    #
     def search_repeated_files
       @repetitions = first_run
       return repetitions if repetitions.empty? || @fast
@@ -33,28 +27,26 @@ module Crf
 
     private
 
-    #
-    # Gets all file paths in the given directory and subdirectories.
-    #
-    def all_files(path)
-      @paths = []
-      Dir["#{path.chomp('/')}/**/*"].each { |p| paths << p.freeze if file?(p) }
-      paths
+    def all_files
+      @files = []
+      paths.each do |path|
+        Dir["#{path.chomp('/')}/**/*"].each do |file_path|
+          @files << file_path.freeze if file?(file_path) && !@files.include?(file_path)
+        end
+      end
+      @files
     end
 
-    #
-    # Checks if the file is not a symlink or a directory.
-    #
     def file?(path)
       !File.directory?(path) && !File.symlink?(path)
     end
 
-    #
+    ##
     # This looks for the files with the same size only
     #
     def first_run
       repetitions_list = Crf::RepetitionsList.new
-      all_files(path).each do |file_path|
+      all_files.each do |file_path|
         repetitions_list.add(file_identifier(file_path).freeze, file_path)
       end
       repetitions_list.repetitions
@@ -64,7 +56,7 @@ module Crf
       File.size(path).to_s
     end
 
-    #
+    ##
     # After finding files with the same size, perform a deeper analysis of those
     #
     def second_run(repetitions)

data/lib/crf/interactive_finder.rb

@@ -2,17 +2,9 @@ require 'crf/repetitions_list'
 require 'ruby-progressbar'
 
 module Crf
-  #
-  # This class finds the paths of all the repeated files inside the path passed as argument.
-  # All files repeated have the same file_identifier and file_hash.
-  #
   class InteractiveFinder < Crf::Finder
-    #
-    # Method that looks for the repeated files in the path specified when the object was created
-    # showing progress bars.
-    #
     def search_repeated_files
-      all_paths = all_files(path)
+      all_paths = all_files
       progressbar = ProgressBar.create(title: 'First run', total: all_paths.count,
                                        format: '%t: %c/%C %a |%B| %%%P')
       rep_list = first_run(progressbar)
@@ -22,12 +14,9 @@ module Crf
 
     private
 
-    #
-    # This looks for the files with the same size only
-    #
     def first_run(progressbar)
       repetitions_list = Crf::RepetitionsList.new
-      all_files(path).each do |file_path|
+      all_files.each do |file_path|
         repetitions_list.add(file_identifier(file_path), file_path)
         progressbar.increment
       end

data/lib/crf/interactive_remover.rb

@@ -1,15 +1,8 @@
 require 'colorize'
+require 'pp'
 
 module Crf
-  #
-  # This is a subclass of Crf::Remover.
-  # It asks the user if he/she wants to remove each file.
-  #
   class InteractiveRemover < Crf::Remover
-    #
-    # This method asks the user if he/she wants to delete each of the files contained in the values
-    # of the repetitions hash.
-    #
     def remove
       saved = 0
       repetitions.each_value do |paths|
@@ -25,13 +18,13 @@ module Crf
 
     def print_all_paths(paths)
       STDOUT.puts 'Found this repetitions:'.green
-      STDOUT.puts paths.to_s.green
+      STDOUT.puts paths.pretty_inspect.green
     end
 
     def remove_confirmation(path)
       STDOUT.print "Do you want to delete the file #{path}? [y/n] ".yellow
       logger.write "Asking to remove #{path}"
-      answer = STDIN.gets.chomp
+      answer = $stdin.gets.chomp
       logger.write "User input: #{answer}"
       if answer == 'y'
         STDOUT.puts "Removed #{path}".red

data/lib/crf/logger.rb

@@ -1,23 +1,19 @@
 require 'logger'
 
 module Crf
-  #
-  # This class is a wrapper of the Logger class, it hanldes the creation and sets the configuration
-  #
   class Logger
-    #
+    ##
     # Creates the logger with the configurations in the path provided or in the current directory
     #
+    # @param path [String] path where the logger is or will be created.
+    #
     def initialize(path = 'crf.log')
       @logger = ::Logger.new(path, File::CREAT)
       configurate_logger
     end
 
-    #
-    # Wrapper of the Logger info method
-    #
-    def write(msg)
-      @logger.info msg
+    def write(message)
+      @logger.info message
     end
 
     private
@@ -25,8 +21,8 @@ module Crf
     def configurate_logger
       @logger.datetime_format = Crf::LOGGER_DATE_TIME_FORMAT
       @logger.progname = Crf::GEM_NAME
-      @logger.formatter = proc do |_severity, datetime, progname, msg|
-        "[#{datetime}] #{progname}: #{msg}\n"
+      @logger.formatter = proc do |_severity, date_time, program_name, message|
+        "[#{date_time}] #{program_name}: #{message}\n"
       end
     end
   end

data/lib/crf/remover.rb

@@ -1,25 +1,21 @@
 module Crf
-  #
-  # This class removes all the repetitions passed as an argument.
-  # It saves the first element of the repetitions and deletes the rest.
-  #
   class Remover
-    #
-    # The repetitions hash and the logger file are accessible from the outside.
-    #
     attr_reader :repetitions, :logger
 
-    #
+    ##
     # This object needs the repeated files obtained with Crf::Finder and the logger object.
     #
+    # @param repetitions [Hash] repetitions found by one of the finder classes.
+    # @param logger [Crf::Logger] logger file which this class will write.
+    #
     def initialize(repetitions, logger)
       @repetitions = repetitions
       @logger = logger
     end
 
-    #
+    ##
     # This method removes all the files contained on each value of the repetitions hash
-    # except the first one. This is done without asking the user for confirmation so be careful.
+    # except the first one. This is done without asking the user for confirmation.
     #
     def remove
       saved = 0
@@ -36,13 +32,11 @@ module Crf
 
     def remove_file(path)
       size = File.size(path)
-      begin
-        File.delete(path)
-        log_removal(path, size)
-        return size
-      rescue
-        return 0
-      end
+      File.delete(path)
+      log_removal(path, size)
+      return size
+    rescue
+      return 0
     end
 
     def log_removal(path, size)

data/lib/crf/repetitions_list.rb

@@ -1,27 +1,20 @@
 module Crf
-  #
-  # This is the data structure used to get the repeated files.
-  #
   class RepetitionsList
-    #
-    # The uniques and repetitions hashes can be accessed from the outside, along with the
-    # total repetitions count.
-    #
     attr_reader :uniques, :repetitions, :total_repetitions
 
-    #
-    # Creates the RepetitionsList object with everything it needs.
-    #
     def initialize
       @uniques = {}
       @repetitions = {}
       @total_repetitions = 0
     end
 
-    #
+    ##
     # Adds an element to one of each hashes. If the value is repeated, then it erases it from
     # uniques and adds it in the repetitions hash along with the duplicate.
     #
+    # @param key result of the function that identifies the file
+    # @param value [String] path of the file
+    #
     def add(key, value)
       if repetitions.key?(key)
         repetitions[key] << value

data/lib/crf/version.rb

@@ -1,6 +1,3 @@
-#
-# Crf version
-#
 module Crf
-  VERSION = '0.0.7'.freeze
+  VERSION = '0.0.8'.freeze
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: crf
 version: !ruby/object:Gem::Version
-  version: 0.0.7
+  version: 0.0.8
 platform: ruby
 authors:
 - Alejandro Bezdjian
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-01-24 00:00:00.000000000 Z
+date: 2016-09-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: colorize
@@ -74,30 +74,42 @@ dependencies:
   name: byebug
   requirement: !ruby/object:Gem::Requirement
     requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '9.0'
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 9.0.5
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '9.0'
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 9.0.5
 - !ruby/object:Gem::Dependency
   name: rubocop
   requirement: !ruby/object:Gem::Requirement
     requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.38'
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.37.2
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.38'
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.37.2
 description: Library that looks for exact duplicated files in a directory.
 email: alebezdjian@gmail.com
 executables:
@@ -106,6 +118,7 @@ extensions: []
 extra_rdoc_files: []
 files:
 - ".gitignore"
+- ".rspec"
 - ".rubocop.yml"
 - ".travis.yml"
 - Gemfile
@@ -143,9 +156,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.5.1
+rubygems_version: 2.4.5.1
 signing_key: 
 specification_version: 4
 summary: Look for exact duplicated files.
 test_files: []
-has_rdoc: