libis-format 0.9.1 → 0.9.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d75c38cec7b9bcbe8ec92de39c229cb4bc3ea8f1
-  data.tar.gz: bd29485e129a54d667414d448d3efd34cc4cb0f6
+  metadata.gz: d3cd5631909a7f886d895de734e4f3ca542dac4b
+  data.tar.gz: 80b1a19961bbbd0f1db1bbcfc06f20674b115417
 SHA512:
-  metadata.gz: da9c536bd82e23ed9aec7e156dbc8f6119f681bb86a8c94c13e59b7995981b925bf93701bad084ffbc7a3b93b822e2d26004d45dc262baabc4bfd3b08657aefa
-  data.tar.gz: b42eaa83bcd737f87914c8d3b30bf02e65094f6e6a8bcdff03fe343f62546f12774880030e0a16f702d40dbb7d6e5540ba4a5e6002f2f6d52fa58db32397fbe4
+  metadata.gz: e7dda44c769f213e908745eba06e65d85252c8fbe9b54dd16b61ea9e4ceea2bbaaecf38c2db12c28e0186331fa36ec0a8303f2a5505610b881c90cc8e5e3ee71
+  data.tar.gz: 2b3df006e9d35f4dc1dd4543790d301a600bb80fabd059bd80b78b1c63f0111430bb3c308f3ef8976a2aa3329b4f1a07bffda79832a5f17a1c8503f989203d28

data/.gitignore

@@ -6,10 +6,13 @@
 /doc/
 /pkg/
 /spec/reports/
+/spec/work/
 /tmp/
 *.bundle
 *.so
 *.o
 *.a
 mkmf.log
-/.idea/
+/.idea/
+*.pyc
+tools/pdf/

data/.travis.yml

@@ -1,21 +1,17 @@
 language: ruby
+sudo: false
 cache: bundler
 rvm:
-  - 1.9.3
   - 2.1.0
   - 2.2.0
   - ruby-head
-  - jruby-19mode
+  - jruby-9.0.1.0
 jdk:
   - openjdk7
   - oraclejdk7
   - oraclejdk8
 matrix:
   exclude:
-    - rvm: 1.9.3
-      jdk: oraclejdk7
-    - rvm: 1.9.3
-      jdk: oraclejdk8
     - rvm: 2.1.0
       jdk: oraclejdk7
     - rvm: 2.1.0
@@ -28,9 +24,18 @@ matrix:
       jdk: oraclejdk7
     - rvm: ruby-head
       jdk: oraclejdk8
+  allow_failures:
+    - rvm: 2.1.0
+    - rvm: 2.2.0
+    - rvm: ruby-head
+    - rvm: jruby-9.0.1.0
 branches:
   only:
       - master
-before_install:
-  - sudo apt-get update -qq
-  - sudo apt-get install -y python2.7
+addons:
+  apt:
+    packages:
+      - python2.7
+      - imagemagick
+      - libreoffice
+      - ghostscript

data/README.md

@@ -1,3 +1,4 @@
+[![Gem Version](https://badge.fury.io/rb/libis-format.svg)](http://badge.fury.io/rb/libis-format)
 [![Build Status](https://travis-ci.org/Kris-LIBIS/LIBIS_Format.svg?branch=master)](https://travis-ci.org/Kris-LIBIS/LIBIS_Format)
 [![Coverage Status](https://img.shields.io/coveralls/Kris-LIBIS/LIBIS_Format.svg)](https://coveralls.io/r/Kris-LIBIS/LIBIS_Format)
 [![Dependency Status](https://gemnasium.com/Kris-LIBIS/LIBIS_Format.svg)](https://gemnasium.com/Kris-LIBIS/LIBIS_Format)

data/bin/pdf_copy

@@ -0,0 +1,13 @@
+#!/usr/bin/env ruby
+
+require 'libis-format'
+require 'libis-tools'
+
+::Libis::Tools::Config.logger.level = ::Logger::Severity::WARN
+
+source = ARGV.shift
+target = ARGV.shift
+options = ARGV
+
+::Libis::Format::PdfCopy.run source, target, options
+

data/data/PDFA_def.ps

@@ -1,23 +1,31 @@
 %!
-% $Id$
 % This is a sample prefix file for creating a PDF/A document.
 % Feel free to modify entries marked with "Customize".
-
 % This assumes an ICC profile to reside in the file (ISO Coated sb.icc),
 % unless the user modifies the corresponding line below.
 
 % Define entries in the document Info dictionary :
-
-/ICCProfile (ISOcoated.icc)   % Customize.
+/ICCProfile ([** Fill in ICC profile location **])
 def
 
-[ /Title (Title)                  % Customize.
+[ /Title (Title)
   /DOCINFO pdfmark
 
 % Define an ICC profile :
 
 [/_objdef {icc_PDFA} /type /stream /OBJ pdfmark
-[{icc_PDFA} <</N systemdict /ProcessColorModel get /DeviceGray eq {1} {4} ifelse >> /PUT pdfmark
+[{icc_PDFA}
+<<
+  /N currentpagedevice /ProcessColorModel known {
+    currentpagedevice /ProcessColorModel get dup /DeviceGray eq
+    {pop 1} {
+      /DeviceRGB eq
+      {3}{4} ifelse
+    } ifelse
+  } {
+    (ERROR, unable to determine ProcessColorModel) == flush
+  } ifelse
+>> /PUT pdfmark
 [{icc_PDFA} ICCProfile (r) file /PUT pdfmark
 
 % Define the output intent dictionary :
@@ -27,6 +35,6 @@ def
   /Type /OutputIntent             % Must be so (the standard requires).
   /S /GTS_PDFA1                   % Must be so (the standard requires).
   /DestOutputProfile {icc_PDFA}            % Must be so (see above).
-  /OutputConditionIdentifier (CGATS TR001)      % Customize
+  /OutputConditionIdentifier ([** Fill in ICC reference name **])
 >> /PUT pdfmark
 [{Catalog} <</OutputIntents [ {OutputIntent_PDFA} ]>> /PUT pdfmark

data/data/types.yml

@@ -13,12 +13,12 @@ IMAGE:
     MIME: image/tiff
     EXTENSIONS: tif,tiff
 
-  JPEG2000:
+  JP2:
     NAME: Joint Photographic Experts Group 2000 (JPEG 2000)
     MIME: image/jp2
     EXTENSIONS: jp2
 
-  JPEG:
+  JPG:
     NAME: Joint Photographic Experts Group (JPEG)
     MIME: image/jpeg
     EXTENSIONS: jpg,jpe,jpeg
@@ -120,7 +120,7 @@ VIDEO:
     MIME: video/x-flv
     EXTENSIONS: flv
 
-DOCUMENT:
+TEXT:
 
   TXT:
     NAME: Unformatted text
@@ -150,28 +150,6 @@ DOCUMENT:
     MIME: application/vnd.openxmlformats-officedocument.wordprocessingml.document
     EXTENSIONS: docx
 
-  MSXLS:
-    NAME: Microsoft Excel Spreadsheet (XLS)
-    MIME: application/vnd.ms-excel,application/msexcel
-    EXTENSIONS: xls
-
-  MSXLSX:
-    NAME: Microsoft Excel OpenXML Spreadheet (XSLX)
-    PUID: fido-fmt/189.xl
-    MIME: application/vnd.openxmlformats-officedocument.spreadsheetml.sheet
-    EXTENSIONS: xslx
-
-  MSPPT:
-    NAME: Microsoft Powerpoint Presentation (PPT)
-    MIME: application/vnd.ms-powerpoint,application/mspowerpoint
-    EXTENSIONS: ppt
-
-  MSPPTX:
-    NAME: Microsoft Powerpoint OpenXML Presentation (PPTX)
-    PUID: fido-fmt/189.ppt
-    MIME: application/vnd.openxmlformats-officedocument.presentationml.presentation
-    EXTENSIONS: pptx
-
   PDF:
     NAME: Adobe Portable Document Format (PDF)
     PUID: fmt/14 fmt/15 fmt/16 fmt/17 fmt/18 fmt/19 fmt/20 fmt/276
@@ -203,6 +181,32 @@ DOCUMENT:
     PUID: fmt/101
     EXTENSIONS: xml
 
+TABULAR:
+
+  MSXLS:
+    NAME: Microsoft Excel Spreadsheet (XLS)
+    MIME: application/vnd.ms-excel,application/msexcel
+    EXTENSIONS: xls
+
+  MSXLSX:
+    NAME: Microsoft Excel OpenXML Spreadheet (XSLX)
+    PUID: fido-fmt/189.xl
+    MIME: application/vnd.openxmlformats-officedocument.spreadsheetml.sheet
+    EXTENSIONS: xslx
+
+PRESENTATION:
+
+  MSPPT:
+    NAME: Microsoft Powerpoint Presentation (PPT)
+    MIME: application/vnd.ms-powerpoint,application/mspowerpoint
+    EXTENSIONS: ppt
+
+  MSPPTX:
+    NAME: Microsoft Powerpoint OpenXML Presentation (PPTX)
+    PUID: fido-fmt/189.ppt
+    MIME: application/vnd.openxmlformats-officedocument.presentationml.presentation
+    EXTENSIONS: pptx
+
 ARCHIVE:
 
   EAD:

data/lib/libis/format/config.rb

@@ -0,0 +1,35 @@
+# encoding: utf-8
+require 'os'
+
+module Libis
+  module Format
+
+    # noinspection RubyConstantNamingConvention
+    Config = ::Libis::Tools::Config
+
+    Config[:converter_chain_max_level] = 8
+
+    Config[:java_path] = 'java'
+    Config[:soffice_path] = 'soffice'
+    Config[:ghostscript_path] = 'gs'
+    # Config[:pdfa_path] =
+    #     File.absolute_path(
+    #         File.join(
+    #             File.dirname(__FILE__), '..', '..', '..', 'tools', 'pdf', 'pdfa', 'pdfa'
+    #         )
+    #     )
+    Config[:droid_path] =
+        File.absolute_path(
+            File.join(
+                File.dirname(__FILE__), '..', '..', '..', 'tools', 'droid', OS.windows? ? 'droid.bat' : 'droid.sh'
+            )
+        )
+    Config[:fido_path] =
+        File.absolute_path(
+            File.join(
+                File.dirname(__FILE__), '..', '..', '..', 'tools', 'fido', OS.windows? ? 'fido.bat' : 'fido.sh'
+            )
+        )
+
+  end
+end

data/lib/libis/format/converter/base.rb

@@ -1,7 +1,7 @@
 # coding: utf-8
 
 ### require 'tools/string'
-
+require 'tmpdir'
 require 'libis/tools/logger'
 require 'libis/format/type_database'
 
@@ -14,37 +14,36 @@ module Libis
       class Base
         include Libis::Tools::Logger
 
-        def input_types
-          raise RuntimeError, 'Method #input_types needs to be overridden in converter'
-        end
-
-        protected
+        attr_reader :options, :flags
 
-        def output_types
-          raise RuntimeError, 'Method #output_types needs to be overridden in converter'
+        def initialize
+          @options = {}
+          @flags = {}
         end
 
-        attr_accessor :source, :options, :flags
-
-        def init(_)
-          raise RuntimeError, 'Method #init should be implemented in converter'
+        def convert(source, target, format, opts = {})
+          unless File.exist? source
+            error "Cannot find file '#{source}'."
+            return nil
+          end
+          @options.merge!(opts[:options]) if opts[:options]
+          @flags.merge!(opts[:flags]) if opts[:flags]
         end
 
-        def do_convert(_, _)
-          raise RuntimeError, 'Method #do_convert should be implemented in converter'
+        def self.input_types(_ = nil)
+          raise RuntimeError, 'Method #input_types needs to be overridden in converter'
         end
 
-        public
-
-        def initialize( source = nil, options = {}, flags = {} )
-          @source = source
-          @options = options ? options : {}
-          @flags = flags ? flags : {}
-          init(source.to_s rescue nil)
+        def self.output_types(_ = nil)
+          raise RuntimeError, 'Method #output_types needs to be overridden in converter'
         end
 
-        def convert(target, format = nil)
-          do_convert(target, format)
+        def using_temp(target)
+          tempfile = File.join(Dir.tmpdir, Dir::Tmpname.make_tmpname(['convert', File.extname(target)], File.basename(target, '.*')))
+          result = yield tempfile
+          return nil unless result
+          FileUtils.move result, target
+          target
         end
 
         def Base.inherited( klass )
@@ -54,7 +53,7 @@ module Libis
           class << self
 
             def conversions
-              input_types.inject({}) do |input_type, hash|
+              input_types.inject({}) do |hash, input_type|
                 hash[input_type] = output_types
                 hash
               end
@@ -92,10 +91,8 @@ module Libis
 
           end
 
-
         end
 
-
       end
 
     end

data/lib/libis/format/converter/chain.rb

@@ -1,6 +1,7 @@
 # coding: utf-8
 
 require 'fileutils'
+require 'deep_dive'
 
 require 'libis/tools/logger'
 require 'libis/format/type_database'
@@ -11,66 +12,164 @@ module Libis
 
       class Chain
         include ::Libis::Tools::Logger
+        include DeepDive
 
-        def initialize(converter_chain)
-          @converter_chain = converter_chain
+        def initialize(source_format, target_format, operations = {})
+          @source_format = source_format
+          @target_format = target_format
+          @operations = operations
+          @converter_chain = []
+        end
+
+        # @param [Libis::Format::Converter::Base.class] converter
+        # @return [Array[Hash]]
+        def append(converter)
+          return [] unless converter
+          valid_chain_nodes(converter).map do |node|
+            self.ddup.add_chain_node(node)
+          end.compact
+        end
+
+        def closed?
+          !@converter_chain.empty? &&
+              @converter_chain.first[:input] == @source_format &&
+              @converter_chain.last[:output] == @target_format
+        end
+
+        def valid?
+          closed? && apply_operations
         end
 
         def to_array
           @converter_chain
         end
 
-        def convert(src_file, target_file, operations = [])
+        def size
+          @converter_chain.size
+        end
 
-          chain = @converter_chain.clone
+        alias_method :length, :size
+
+        def to_s
+          # nodes_string = @converter_chain.map do |node|
+          #   node_name = node[:converter].name.gsub(/^.*::/,'')
+          #   node_operations = '(' + node[:operations].map do |operation|
+          #     op = "#{operation[:method]}:#{operation[:argument]}"
+          #     op
+          #   end.join(', ') + ')' rescue ''
+          #   node_string = "#{node_name}#{node_operations}->-#{node[:output]}"
+          #   node_string
+          # end.join('->-')
+          # "#{@source_format}->-#{nodes_string}"
+          "#{@source_format}->-#{@converter_chain.map do |node|
+            "#{node[:converter].name.gsub(/^.*::/,'')}#{node[:operations] ?
+                "(#{node[:operations].each { |operation| "#{operation[:method]}:#{operation[:argument]}" }.join(',')})" :
+                ''
+            }->-#{node[:output]}"
+          end.join('->-')}"
+        end
 
-          my_operations = {}
+        def convert(src_file, target_file)
 
-          # sanity check: check if the required operations are supported by at least one converter in the chain
-          operations.each do |k,v|
-            method = k.to_s.downcase.to_sym
-            chain_element = @converter_chain.reverse.detect { |c| c[:converter].new.respond_to? method }
-            if chain_element
-              my_operations[chain_element[:converter]] ||= {}
-              my_operations[chain_element[:converter]][method] = v
-            else
-              error "No converter in the converter chain supports '#{method.to_s}'. Continuing conversion without this operation."
-            end
+          unless valid?
+            error 'Converter chain is not valid'
+            return nil
           end
 
           temp_files = []
 
           # noinspection RubyParenthesesAroundConditionInspection
-          while (chain_element = chain.shift)
+          result = @converter_chain.each_with_index do |node, i|
 
-            target_type = chain_element[:target]
-            converter_class = chain_element[:converter]
-            converter = converter_class.new(src_file)
+            target_type = node[:output]
+            converter_class = node[:converter]
+            converter = converter_class.new
 
-            my_operations[converter_class].each do |k,v|
-              converter.send k, v
+            node[:operations].each do |operation|
+              converter.send operation[:method], operation[:argument]
             end
 
             target = target_file
 
-            unless chain.empty?
-              target += '.temp.' + TypeDatabase.instance.type2ext(target_type)
-              target += '.' + TypeDatabase.instance.type2ext(target_type) while File.exist? target
+            if i < size
+              target += ".temp.#{TypeDatabase.type_extentions(target_type).first}"
+              target += ".#{TypeDatabase.type_extentions(target_type).first}" while File.exist? target
               temp_files << target
             end
 
             FileUtils.mkdir_p File.dirname(target)
 
-            converter.convert(target, target_type)
+            src_file = converter.convert(src_file, target, target_type)
 
-            src_file = target
+            break :failed unless src_file
 
           end
 
           temp_files.each do |f|
-            File.delete(f)
+            FileUtils.rm(f, force: true)
+          end
+
+          result == :failed ? nil : target_file
+
+        end
+
+        def valid_chain_nodes(converter)
+          return [] if closed?
+          source_format = @converter_chain.last[:output] rescue @source_format
+          nodes = []
+          if converter.input_types.include? source_format
+            converter.output_types(source_format).each do |format|
+              node = {converter: converter, input: source_format, output: format}
+              next if node_exists?(node)
+              nodes << node
+            end
           end
+          nodes
+        end
 
+        def add_chain_node(node = {})
+          return nil if closed?
+          last_converter = @converter_chain.last
+          source_format = last_converter ? last_converter[:output] : @source_format
+          node[:input] ||= source_format
+          return nil unless node[:input] == source_format
+          return nil unless node[:output] && node[:converter].output_types(source_format).include?(node[:output])
+          return nil unless node[:converter].input_types.include? source_format
+          return nil if node_exists?(node)
+          @converter_chain << node
+          # debug "Chain: #{self}"
+          self
+        end
+
+        def apply_operations
+          return false unless closed?
+          temp_chain = @converter_chain.reverse.ddup
+          applied = true
+          operations = @operations.ddup
+          while (operation = operations.shift)
+            method = operation.first.to_s.downcase.to_sym
+            applied &&= :found == temp_chain.each do |node|
+              next unless node[:converter].instance_methods.include?(method)
+              node[:operations] ||= []
+              node[:operations] << {method: method, argument: operation.last}
+              break :found
+            end
+          end
+          if applied && operations.empty?
+            @converter_chain = temp_chain.reverse
+            @operations.clear
+            return true
+          end
+          false
+        end
+
+
+        private
+
+        def node_exists?(node)
+          @converter_chain.detect do |n|
+            n[:converter] == node[:converter] && n[:input] == node[:input] && n[:output] == node[:output]
+          end
         end
 
       end