assembly-objectfile 1.13.0 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8c85bb6c05fe0d46e9b2f37e90710043efce16322c5d58e0b0a9578fd21ece77
-  data.tar.gz: c7e3ac5e3f0a0aa2124f8ce3d6587de2d02923c733ee94b1c299ae171101f5a5
+  metadata.gz: c3c456d372ab0804ebcaad06ac2b8695376f9ffe57ccb8a23b8a57d711df5561
+  data.tar.gz: 1797606727f148ee128dedc102ad68480fbd9f55983c318b91c8f5f52a546a9b
 SHA512:
-  metadata.gz: fbad307e9a103af61b4a6b05b35bb6b98953774a2c9bd32cb1404f1a9ff6e3404fe78cfc48f9061bfffaa23c3610f2a6f19df5e08975a20c6ee387073cc1010d
-  data.tar.gz: 717e174ccf88a94880f4a832b2d281b351774299fb1273a96b4a1bb942e360cd20b0ba3ac68b7ad4f113e3bd3c78652c7670efcf7bcf8f9bcf068972ba5cd42d
+  metadata.gz: c927cb2cffef41a699c0f2e1f0fa0d8e6d4b1cc92c9fdb26b80ca5c9e851d7de0a9b1204cc4a4290213a9d7a642c2d3233db56396efb3b626cdd4d6e7b3bfdda
+  data.tar.gz: 473f2ead345feaf0afd518da399751a86f526f27b19dd0bfd75fb55354bd055d00ccbbb06a96ca905f2e0de6fcdbf25d4e42e8da23fc167225afb8a9958038e3

data/.gitignore

@@ -10,4 +10,3 @@ log/*
 pkg/*
 tags
 tmp
-Gemfile.lock

data/.rubocop.yml

@@ -14,7 +14,7 @@ Metrics/BlockLength:
     - 'spec/**/*.rb'
     - '**/*.gemspec'
 
-Gemspec/DateAssignment: # (new in 1.10)
+Gemspec/DeprecatedAttributeAssignment: # (new in 1.10)
   Enabled: true
 
 Layout/SpaceAroundMethodCallOperator:

data/.rubocop_todo.yml

@@ -1,6 +1,6 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2022-02-28 21:11:02 UTC using RuboCop version 1.25.1.
+# on 2022-07-07 21:26:21 UTC using RuboCop version 1.31.0.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
@@ -11,41 +11,10 @@ Lint/UselessAssignment:
   Exclude:
     - 'config/boot.rb'
 
-# Offense count: 3
-# Configuration parameters: IgnoredMethods, CountRepeatedAttributes.
-Metrics/AbcSize:
-  Max: 55
-
-# Offense count: 1
-# Configuration parameters: CountComments, CountAsOne, ExcludedMethods, IgnoredMethods.
-# IgnoredMethods: refine
-Metrics/BlockLength:
-  Max: 27
-
-# Offense count: 2
-# Configuration parameters: IgnoredMethods.
-Metrics/CyclomaticComplexity:
-  Max: 14
-
-# Offense count: 4
-# Configuration parameters: CountComments, CountAsOne, ExcludedMethods, IgnoredMethods.
-Metrics/MethodLength:
-  Max: 31
-
 # Offense count: 1
 # Configuration parameters: CountComments, CountAsOne.
-Metrics/ModuleLength:
-  Max: 120
-
-# Offense count: 1
-# Configuration parameters: CountKeywordArgs, MaxOptionalParameters.
-Metrics/ParameterLists:
-  Max: 12
-
-# Offense count: 2
-# Configuration parameters: IgnoredMethods.
-Metrics/PerceivedComplexity:
-  Max: 15
+Metrics/ClassLength:
+  Max: 122
 
 # Offense count: 1
 # Configuration parameters: ExpectMatchingDefinition, CheckDefinitionPathHierarchy, CheckDefinitionPathHierarchyRoots, Regex, IgnoreExecutableScripts, AllowedAcronyms.
@@ -64,66 +33,31 @@ Naming/FileName:
 Naming/PredicateName:
   Exclude:
     - 'spec/**/*'
-    - 'lib/assembly-objectfile/object_fileable.rb'
-
-# Offense count: 1
-# Configuration parameters: Prefixes.
-# Prefixes: when, with, without
-RSpec/ContextWording:
-  Exclude:
-    - 'spec/content_metadata_spec.rb'
+    - 'lib/assembly-objectfile/object_file.rb'
 
-# Offense count: 32
+# Offense count: 2
 # Configuration parameters: CountAsOne.
 RSpec/ExampleLength:
-  Max: 34
+  Max: 7
 
-# Offense count: 2
+# Offense count: 1
 # Configuration parameters: Include, CustomTransform, IgnoreMethods, SpecSuffixOnly.
 # Include: **/*_spec*rb*, **/spec/**/*
 RSpec/FilePath:
   Exclude:
-    - 'spec/content_metadata_spec.rb'
     - 'spec/object_file_spec.rb'
 
-# Offense count: 74
-# Configuration parameters: AssignmentOnly.
-RSpec/InstanceVariable:
-  Exclude:
-    - 'spec/object_file_spec.rb'
-
-# Offense count: 41
+# Offense count: 8
 RSpec/MultipleExpectations:
-  Max: 29
+  Max: 6
 
-# Offense count: 20
+# Offense count: 3
 RSpec/NestedGroups:
   Max: 4
 
-# Offense count: 2
-RSpec/RepeatedDescription:
-  Exclude:
-    - 'spec/object_file_spec.rb'
-
-# Offense count: 2
-RSpec/RepeatedExample:
-  Exclude:
-    - 'spec/object_file_spec.rb'
-
-# Offense count: 5
-RSpec/RepeatedExampleGroupDescription:
-  Exclude:
-    - 'spec/content_metadata_spec.rb'
-
-# Offense count: 2
-# Cop supports --auto-correct.
-Style/CommentedKeyword:
-  Exclude:
-    - 'lib/assembly-objectfile/content_metadata.rb'
-
-# Offense count: 123
-# Cop supports --auto-correct.
-# Configuration parameters: AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, IgnoredPatterns.
+# Offense count: 24
+# This cop supports safe autocorrection (--autocorrect).
+# Configuration parameters: AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, AllowedPatterns, IgnoredPatterns.
 # URISchemes: http, https
 Layout/LineLength:
-  Max: 277
+  Max: 187

data/Gemfile.lock

@@ -0,0 +1,106 @@
+PATH
+  remote: .
+  specs:
+    assembly-objectfile (2.0.0)
+      activesupport (>= 5.2.0)
+      deprecation
+      mime-types (> 3)
+      mini_exiftool
+      nokogiri
+
+GEM
+  remote: http://rubygems.org/
+  specs:
+    activesupport (7.0.3)
+      concurrent-ruby (~> 1.0, >= 1.0.2)
+      i18n (>= 1.6, < 2)
+      minitest (>= 5.1)
+      tzinfo (~> 2.0)
+    ast (2.4.2)
+    byebug (11.1.3)
+    coderay (1.1.3)
+    concurrent-ruby (1.1.10)
+    deprecation (1.1.0)
+      activesupport
+    diff-lcs (1.5.0)
+    docile (1.4.0)
+    i18n (1.10.0)
+      concurrent-ruby (~> 1.0)
+    json (2.6.2)
+    method_source (1.0.0)
+    mime-types (3.4.1)
+      mime-types-data (~> 3.2015)
+    mime-types-data (3.2022.0105)
+    mini_exiftool (2.10.2)
+    mini_portile2 (2.8.0)
+    minitest (5.16.2)
+    nokogiri (1.13.6)
+      mini_portile2 (~> 2.8.0)
+      racc (~> 1.4)
+    parallel (1.22.1)
+    parser (3.1.2.0)
+      ast (~> 2.4.1)
+    pry (0.13.1)
+      coderay (~> 1.1)
+      method_source (~> 1.0)
+    pry-byebug (3.9.0)
+      byebug (~> 11.0)
+      pry (~> 0.13.0)
+    racc (1.6.0)
+    rainbow (3.1.1)
+    rake (13.0.6)
+    regexp_parser (2.5.0)
+    rexml (3.2.5)
+    rspec (3.11.0)
+      rspec-core (~> 3.11.0)
+      rspec-expectations (~> 3.11.0)
+      rspec-mocks (~> 3.11.0)
+    rspec-core (3.11.0)
+      rspec-support (~> 3.11.0)
+    rspec-expectations (3.11.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.11.0)
+    rspec-mocks (3.11.1)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.11.0)
+    rspec-support (3.11.0)
+    rubocop (1.31.0)
+      parallel (~> 1.10)
+      parser (>= 3.1.0.0)
+      rainbow (>= 2.2.2, < 4.0)
+      regexp_parser (>= 1.8, < 3.0)
+      rexml (>= 3.2.5, < 4.0)
+      rubocop-ast (>= 1.18.0, < 2.0)
+      ruby-progressbar (~> 1.7)
+      unicode-display_width (>= 1.4.0, < 3.0)
+    rubocop-ast (1.18.0)
+      parser (>= 3.1.1.0)
+    rubocop-rspec (2.11.1)
+      rubocop (~> 1.19)
+    ruby-progressbar (1.11.0)
+    simplecov (0.21.2)
+      docile (~> 1.1)
+      simplecov-html (~> 0.11)
+      simplecov_json_formatter (~> 0.1)
+    simplecov-html (0.12.3)
+    simplecov_json_formatter (0.1.4)
+    tzinfo (2.0.4)
+      concurrent-ruby (~> 1.0)
+    unicode-display_width (2.2.0)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  assembly-objectfile!
+  byebug
+  json
+  pry-byebug
+  rake
+  rspec (~> 3.0)
+  rubocop (~> 1.25)
+  rubocop-rspec
+  simplecov
+
+BUNDLED WITH
+   2.3.4

data/assembly-objectfile.gemspec

@@ -16,7 +16,6 @@ Gem::Specification.new do |s|
   s.metadata['rubygems_mfa_required'] = 'true'
 
   s.files         = `git ls-files`.split("\n")
-  s.test_files    = `git ls-files -- spec/*`.split("\n")
   s.bindir        = 'exe'
   s.executables   = s.files.grep(%r{^exe/}) { |f| File.basename(f) }
   s.require_paths = ['lib']
@@ -25,13 +24,12 @@ Gem::Specification.new do |s|
 
   s.add_dependency 'activesupport', '>= 5.2.0'
   s.add_dependency 'deprecation'
-  s.add_dependency 'dry-struct', '~> 1.0'
-  s.add_dependency 'dry-types', '~> 1.1'
   s.add_dependency 'mime-types', '> 3'
   s.add_dependency 'mini_exiftool'
   s.add_dependency 'nokogiri'
 
   s.add_development_dependency 'json'
+  s.add_development_dependency 'pry-byebug'
   s.add_development_dependency 'rake'
   s.add_development_dependency 'rspec', '~> 3.0'
   s.add_development_dependency 'rubocop', '~> 1.25'

data/lib/assembly-objectfile/object_file.rb

@@ -1,14 +1,16 @@
 # frozen_string_literal: true
 
+require 'mini_exiftool'
+require 'mime/types'
+require 'active_support/core_ext/object/blank'
+
 module Assembly
   # This class contains generic methods to operate on any file.
   class ObjectFile
-    include Assembly::ObjectFileable
-
     # Class level method that given an array of strings, return the longest common initial path.  Useful for removing a common path from a set of filenames when producing content metadata
     #
     # @param [Array] strings Array of filenames with paths to operate on
-    # @return [String] Common part of path of filenames passed in
+    # @return [String] longest common initial part of path of filenames passed in
     #
     # Example:
     # puts Assembly::ObjectFile.common_prefix(['/Users/peter/00/test.tif','/Users/peter/05/test.jp2'])  # '/Users/peter/0'
@@ -25,5 +27,253 @@ module Assembly
         "#{common_prefix.split('/')[0..-2].join('/')}/" # if it was, then return the common prefix directly
       end
     end
+
+    attr_accessor :file_attributes, :label, :path, :provider_md5, :provider_sha1, :relative_path, :mime_type_order
+
+    VALID_MIMETYPE_METHODS = %i[override exif file extension].freeze
+
+    # @param [String] path full path to the file to be worked with
+    # @param [Hash<Symbol => Object>] params options used during content metadata generation
+    # @option params [Hash<Symbol => ['yes', 'no']>] :file_attributes e.g. {:preserve=>'yes',:shelve=>'no',:publish=>'no'}, defaults pulled from mimetype
+    # @option params [String] :label a resource label (files bundlded together will just get the first file's label attribute if set)
+    # @option params [String] :provider_md5 pre-computed MD5 checksum
+    # @option params [String] :provider_sha1 pre-computed SHA1 checksum
+    # @option params [String] :relative_path if you want the file ids in the content metadata it can be set, otherwise content metadata will get the full path
+    # @option params [Array] :mime_type_order can be set to the order in which you want mimetypes to be determined
+    #                                          options are :override (from manual overide mapping if exists), :exif (from exif if exists),
+    #                                                      :extension (from file extension), and :file (from unix file system command)
+    #                                          the default is defined in the private `default_mime_type_order` method but you can override to set your own order
+    # @example
+    #   Assembly::ObjectFile.new('/input/path_to_file.tif')
+    def initialize(path, params = {})
+      @path = path
+      @label = params[:label]
+      @file_attributes = params[:file_attributes]
+      @relative_path = params[:relative_path]
+      @provider_md5 = params[:provider_md5]
+      @provider_sha1 = params[:provider_sha1]
+      @mime_type_order = params[:mime_type_order] || default_mime_type_order
+    end
+
+    # @return [String] base filename
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.filename # "path_to_file.tif"
+    def filename
+      File.basename(path)
+    end
+
+    # @return [String] base directory
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.dirname # "/input"
+    def dirname
+      File.dirname(path)
+    end
+
+    # @return [String] filename extension
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.ext # ".tif"
+    def ext
+      File.extname(path)
+    end
+
+    # @return [String] base filename without extension
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.filename # "path_to_file"
+    def filename_without_ext
+      File.basename(path, ext)
+    end
+
+    # @return [MiniExiftool] exif information stored as a hash and an object
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.exif # hash with exif information
+    def exif
+      @exif ||= begin
+        check_for_file
+        MiniExiftool.new(path, replace_invalid_chars: '?')
+      end
+    end
+
+    # Computes md5 checksum or returns cached value
+    # @return [String] md5 checksum
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.md5 # 'XXX123XXX1243XX1243'
+    def md5
+      check_for_file unless @md5
+      @md5 ||= Digest::MD5.file(path).hexdigest
+    end
+
+    # Computes sha1 checksum or return cached value
+    # @return [String] sha1 checksum
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.sha1 # 'XXX123XXX1243XX1243'
+    def sha1
+      check_for_file unless @sha1
+      @sha1 ||= Digest::SHA1.file(path).hexdigest
+    end
+
+    # Returns mimetype information for the current file based on the ordering set in default_mime_type_order
+    #   We stop computing mimetypes as soon as we have a method that returns a value
+    # @return [String] mime type
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.txt')
+    #   puts source_file.mimetype # 'text/plain'
+    def mimetype
+      @mimetype ||= begin
+        check_for_file
+        mimetype = ''
+        mime_type_order.each do |mime_type_method|
+          mimetype = send("#{mime_type_method}_mimetype") if VALID_MIMETYPE_METHODS.include?(mime_type_method)
+          break if mimetype.present?
+        end
+        mimetype
+      end
+    end
+
+    # @return [Symbol] the type of object, could be :application (for PDF or Word, etc), :audio, :image, :message, :model, :multipart, :text or :video
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.object_type # :image
+    def object_type
+      lookup = MIME::Types[mimetype][0]
+      lookup.nil? ? :other : lookup.media_type.to_sym
+    end
+
+    # @return [Boolean] if object is an image
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.image? # true
+    def image?
+      object_type == :image
+    end
+
+    # Examines the input image for validity.  Used to determine if image is a valid and useful image.
+    # If image is not a jp2, also checks if it is jp2able?
+    # @return [Boolean] true if image is valid, false if not.
+    # @example
+    #   source_img = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_img.valid_image? # true
+    def valid_image?
+      return false unless image?
+
+      mimetype == 'image/jp2' || jp2able?
+    end
+
+    # @return [Boolean] true if image has a color profile, false if not.
+    # @example
+    #   source_img = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_img.has_color_profile? # true
+    def has_color_profile?
+      return false unless exif
+
+      exif['profiledescription'] || exif['colorspace'] ? true : false
+    end
+
+    # Examines the input image for validity to create a jp2.  Same as valid_image? but also confirms the existence of a profile description and further restricts mimetypes.
+    # It is used by the assembly robots to decide if a jp2 will be created and is also called before you create a jp2 using assembly-image.
+    # @return [Boolean] true if image should have a jp2 created, false if not.
+    # @example
+    #   source_img = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_img.jp2able? # true
+    def jp2able?
+      return false unless exif
+
+      Assembly::VALID_IMAGE_MIMETYPES.include?(mimetype)
+    end
+
+    # Returns file size information for the current file in bytes.
+    # @return [Integer] file size in bytes
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.filesize # 1345
+    def filesize
+      check_for_file
+      @filesize ||= File.size(path)
+    end
+
+    # Determines if the file exists (and is not a directory)
+    # @return [Boolean] file exists
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.tif')
+    #   puts source_file.file_exists? # true
+    def file_exists?
+      @file_exists ||= (File.exist?(path) && !File.directory?(path))
+    end
+
+    private
+
+    # private method to check for file existence before operating on it
+    def check_for_file
+      raise "input file #{path} does not exist or is a directory" unless file_exists?
+    end
+
+    # prive method defining default preferred ordering of how mimetypes are determined
+    def default_mime_type_order
+      %i[override exif file extension]
+    end
+
+    # Returns mimetype information using the mime-types gem (based on a file extension lookup)
+    # @return [String] mime type for supplied file
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.txt')
+    #   puts source_file.extension_mimetype # 'text/plain'
+    def extension_mimetype
+      @extension_mimetype ||= begin
+        mtype = MIME::Types.type_for(path).first
+        mtype ? mtype.content_type : ''
+      end
+    end
+
+    # Returns mimetype information for the current file based on unix file system command.
+    # @return [String] mime type for supplied file
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.txt')
+    #   puts source_file.file_mimetype # 'text/plain'
+    def file_mimetype
+      @file_mimetype ||= begin
+        check_for_file
+        `file --mime-type "#{path}"`.delete("\n").split(':')[1].strip # first try and get the mimetype from the unix file command
+      end
+    end
+
+    # Returns mimetype information for the current file based on exif data (if available and not a trusted source that we'd rather get from the file system command)
+    # @return [String] mime type for supplied file
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.txt')
+    #   puts source_file.exif_mimetype # 'text/plain'
+    def exif_mimetype
+      @exif_mimetype ||= begin
+        check_for_file
+        prefer_exif = !Assembly::TRUSTED_MIMETYPES.include?(file_mimetype) # if it's not a "trusted" mimetype and there is exif data; get the mimetype from the exif
+        exif.mimetype if exif&.mimetype && prefer_exif
+      end
+    end
+
+    # Returns mimetype information using the manual override mapping (based on a file extension lookup)
+    # @return [String] mime type for supplied file if a mapping exists for the file's extension
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.json')
+    #   puts source_file.override_mimetype # 'application/json'
+    def override_mimetype
+      @override_mimetype ||= Assembly::OVERRIDE_MIMETYPES.fetch(ext.to_sym, '')
+    end
+
+    # @note Uses shell call to "file", only expected to work on unix based systems
+    # @return [String] encoding for supplied file
+    # @example
+    #   source_file = Assembly::ObjectFile.new('/input/path_to_file.txt')
+    #   puts source_file.encoding # 'us-ascii'
+    def encoding
+      @encoding ||= begin
+        check_for_file
+        `file --mime-encoding "#{path}"`.delete("\n").split(':')[1].strip
+      end
+    end
   end
 end

data/lib/assembly-objectfile/version.rb

@@ -4,6 +4,6 @@
 module Assembly
   class ObjectFile
     # Gem version
-    VERSION = '1.13.0'
+    VERSION = '2.0.0'
   end
 end

data/lib/assembly-objectfile.rb

@@ -7,9 +7,6 @@ module Assembly
   # if input image is not one of these mime types, it will not be regarded as a valid image for the purpose of generating a JP2 derivative
   VALID_IMAGE_MIMETYPES = ['image/jpeg', 'image/tiff', 'image/tif', 'image/png'].freeze
 
-  # if input file has one of these extensions in a 3D object, it will get the 3d resource type
-  VALID_THREE_DIMENSION_EXTENTIONS = ['.obj'].freeze
-
   # the list of mimetypes that will be "trusted" by the unix file command; if a mimetype other than one of these is returned
   #  by the file command, then a check will be made to see if exif data exists...if so, the mimetype returned by the exif data will be used
   #  if no exif data exists, then the mimetype returned by the unix file command will be used
@@ -24,7 +21,5 @@ module Assembly
   }.freeze
 end
 
-require 'assembly-objectfile/content_metadata'
-require 'assembly-objectfile/object_fileable'
 require 'assembly-objectfile/object_file'
 require 'assembly-objectfile/version'