pdf_outline_editor 0.1.0-java

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: f4919cc4df58a6af4c74185a140a8845176b9afef34bed77558171361fbb299e
+  data.tar.gz: 3dd0368ac59022b051083b41b48dc48a21944e1d224d570574651d95298697a5
+SHA512:
+  metadata.gz: 9e523afaa754c0b56534e34dd65688416d6e6b109770890f8bc19b0a7fff8ae586c828e9fe35a2245ea0164582ccb7e96d590a2afe3d8b260158cc27bfd5ad5e
+  data.tar.gz: 77fbe6d6d9d8afaf5bd4eb78aa987a857ef1690e99e6a72379eb487820526aa5e47779a25f30f3f3cd09ca63a3c1697517f0330e9053a79cd0505b0a3a445b84

data/README.md

@@ -0,0 +1,51 @@
+# `pdf_outline_editor`
+
+[![Build Status](https://travis-ci.org/healthypackrat/pdf_outline_editor.svg?branch=master)](https://travis-ci.org/healthypackrat/pdf_outline_editor)
+
+This gem provides a command to get/set PDF outlines from a JSON/YAML definition file.
+
+## Requirements
+
+- JRuby
+
+## Installation
+
+```
+$ jgem install pdf_outline_editor
+```
+
+## Usage
+
+First, generate outlines definition file by running `init` command:
+
+```
+$ pdf_outline_editor init -f json > toc.json
+```
+
+or, by running `dump` command with an existing pdf file:
+
+```
+$ pdf_outline_editor dump -f json input.pdf > toc.json
+```
+
+(Edit `toc.json` as you like...)
+
+Then run `load` command to set outlines:
+
+```
+$ pdf_outline_editor load input.pdf toc.json output.pdf
+```
+
+## Development
+
+Run `bin/jruby-ng-server` in a new terminal window for faster loading.
+
+Then run `bin/rspec` to run the specs.
+
+## Contributing
+
+Bug reports and pull requests are welcome on GitHub at <https://github.com/healthypackrat/pdf_outline_editor>.
+
+## License
+
+The gem is available as open source under the terms of the [Apache License, Version 2.0](https://opensource.org/licenses/Apache-2.0).

data/exe/pdf_outline_editor

@@ -0,0 +1,11 @@
+#!/usr/bin/env jruby
+
+require "pdf_outline_editor"
+
+begin
+  runner = PdfOutlineEditor::Runner.new(ARGV)
+  runner.run
+rescue PdfOutlineEditor::Error => error
+  warn error.message
+  exit 1
+end

data/lib/pdf_outline_editor.rb

@@ -0,0 +1,17 @@
+require 'java'
+
+require_relative '../vendor/pdfbox/pdfbox-app-2.0.13.jar'
+
+require 'pdf_outline_editor/version'
+
+require 'pdf_outline_editor/dumper'
+require 'pdf_outline_editor/loader'
+
+require 'pdf_outline_editor/runner'
+require 'pdf_outline_editor/commands/dump_command'
+require 'pdf_outline_editor/commands/init_command'
+require 'pdf_outline_editor/commands/load_command'
+
+module PdfOutlineEditor
+  class Error < StandardError; end
+end

data/lib/pdf_outline_editor/commands/dump_command.rb

@@ -0,0 +1,68 @@
+require 'optparse'
+
+module PdfOutlineEditor
+  module Commands
+    class DumpCommand
+      attr_reader :output_format, :input_pdf_path
+
+      def initialize(argv)
+        @parser = nil
+
+        @output_formats = [:json, :yaml]
+        @output_format = :json
+
+        @input_pdf_path = nil
+
+        parser.parse!(argv)
+
+        handle_args(argv)
+      end
+
+      def run
+        Dumper.open(@input_pdf_path) do |dumper|
+          outlines = dumper.dump
+
+          if outlines
+            puts send("convert_to_#{@output_format}", outlines)
+          end
+        end
+      end
+
+      private
+
+      def parser
+        return @parser if @parser
+
+        @parser = OptionParser.new
+
+        @parser.banner = "Usage: #{File.basename($0)} dump [options] <input-pdf-path>"
+
+        desc = "Output format (default: #{@output_format}; one of #{@output_formats.join(', ')})"
+        @parser.on('-f', '--format=FORMAT', @output_formats, desc) do |value|
+          @output_format = value.to_sym
+        end
+
+        @parser
+      end
+
+      def handle_args(argv)
+        @input_pdf_path = argv.shift
+        raise Error, 'missing input pdf path' unless @input_pdf_path
+      end
+
+      def convert_to_json(outlines)
+        require 'json'
+
+        JSON.pretty_generate(outlines)
+      end
+
+      def convert_to_yaml(outlines)
+        require 'yaml'
+
+        YAML.dump(outlines)
+      end
+
+      Runner.register_command(:dump, self)
+    end
+  end
+end

data/lib/pdf_outline_editor/commands/init_command.rb

@@ -0,0 +1,70 @@
+require 'optparse'
+
+module PdfOutlineEditor
+  module Commands
+    class InitCommand
+      attr_reader :output_format
+
+      def initialize(argv)
+        @parser = nil
+
+        @output_formats = [:json, :yaml]
+        @output_format = :json
+
+        parser.parse!(argv)
+      end
+
+      def run
+        sample_outlines = [
+          {
+            'title' => 'Chapter 1',
+            'page' => 1,
+            'children' => [
+              {
+                'title' => 'Section 1.1',
+                'page' => 2
+              }
+            ]
+          },
+          {
+            'title' => 'Chapter 2',
+            'page' => 3
+          }
+        ]
+
+        puts send("convert_to_#{@output_format}", sample_outlines)
+      end
+
+      private
+
+      def parser
+        return @parser if @parser
+
+        @parser = OptionParser.new
+
+        @parser.banner = "Usage: #{File.basename($0)} init [options]"
+
+        desc = "Output format (default: #{@output_format}; one of #{@output_formats.join(', ')})"
+        @parser.on('-f', '--format=FORMAT', @output_formats, desc) do |value|
+          @output_format = value.to_sym
+        end
+
+        @parser
+      end
+
+      def convert_to_json(outlines)
+        require 'json'
+
+        JSON.pretty_generate(outlines)
+      end
+
+      def convert_to_yaml(outlines)
+        require 'yaml'
+
+        YAML.dump(outlines)
+      end
+
+      Runner.register_command(:init, self)
+    end
+  end
+end

data/lib/pdf_outline_editor/commands/load_command.rb

@@ -0,0 +1,80 @@
+require 'optparse'
+
+module PdfOutlineEditor
+  module Commands
+    class LoadCommand
+      def initialize(argv)
+        @parser = nil
+
+        @input_pdf_path = nil
+        @outlines = nil
+        @output_pdf_path = nil
+
+        parser.parse!(argv)
+
+        handle_args(argv)
+      end
+
+      def run
+        Loader.open(@input_pdf_path) do |loader|
+          loader.load(@outlines)
+          loader.save(@output_pdf_path)
+        end
+      end
+
+      private
+
+      def parser
+        return @parser if @parser
+
+        @parser = OptionParser.new
+
+        @parser.banner = "Usage: #{File.basename($0)} load <input-pdf-path> <input-outlines-path> <output-pdf-path>"
+
+        @parser
+      end
+
+      def handle_args(argv)
+        @input_pdf_path = argv.shift
+        raise Error, 'missing input pdf path' unless @input_pdf_path
+
+        input_outlines_path = argv.shift
+        raise Error, 'missing input outlines path' unless input_outlines_path
+        @outlines = parse_outlines(input_outlines_path)
+
+        @output_pdf_path = argv.shift
+        raise Error, 'missing output pdf path' unless @output_pdf_path
+      end
+
+      def parse_outlines(path)
+        ext = File.extname(path)[1..-1]
+
+        method_name = "parse_#{ext}_outlines"
+
+        if respond_to?(method_name, true)
+          send(method_name, path)
+        else
+          raise Error, "unknown extension: #{ext}"
+        end
+      end
+
+      def parse_json_outlines(path)
+        require 'json'
+
+        JSON.parse(File.read(path))
+      end
+
+      def parse_yaml_outlines(path)
+        require 'yaml'
+
+        YAML.load_file(path)
+      end
+
+      def parse_yml_outlines(path)
+        parse_yaml_outlines(path)
+      end
+
+      Runner.register_command(:load, self)
+    end
+  end
+end

data/lib/pdf_outline_editor/dumper.rb

@@ -0,0 +1,74 @@
+module PdfOutlineEditor
+  class Dumper
+    java_import org.apache.pdfbox.pdmodel.PDDocument
+
+    JavaFile = java.io.File
+
+    def self.open(input_pdf_path)
+      dumper = new(input_pdf_path)
+
+      begin
+        yield dumper
+      ensure
+        dumper.close
+      end
+    end
+
+    attr_reader :closed
+
+    def initialize(input_pdf_path)
+      begin
+        @doc = PDDocument.load(JavaFile.new(input_pdf_path))
+      rescue
+        raise Error, $!.message
+      end
+
+      @pages = @doc.pages
+
+      @closed = false
+    end
+
+    def dump
+      root_outline = @doc.document_catalog.document_outline
+
+      if root_outline
+        traverse(root_outline)
+      else
+        nil
+      end
+    end
+
+    def close
+      unless @closed
+        @doc.close
+        @closed = true
+      end
+    end
+
+    private
+
+    def traverse(outline)
+      outline_items = []
+
+      current = outline.first_child
+
+      while current
+        outline_item = {}
+
+        outline_item['title'] = current.title
+
+        page = current.find_destination_page(@doc)
+        outline_item['page'] = @pages.index_of(page) + 1
+
+        children = traverse(current)
+        outline_item['children'] = children unless children.empty?
+
+        outline_items << outline_item
+
+        current = current.next_sibling
+      end
+
+      outline_items
+    end
+  end
+end

data/lib/pdf_outline_editor/loader.rb

@@ -0,0 +1,85 @@
+module PdfOutlineEditor
+  class Loader
+    java_import org.apache.pdfbox.pdmodel.PDDocument
+    java_import org.apache.pdfbox.pdmodel.interactive.documentnavigation.destination.PDPageXYZDestination
+    java_import org.apache.pdfbox.pdmodel.interactive.documentnavigation.outline.PDDocumentOutline
+    java_import org.apache.pdfbox.pdmodel.interactive.documentnavigation.outline.PDOutlineItem
+
+    JavaFile = java.io.File
+
+    def self.open(input_pdf_path)
+      loader = new(input_pdf_path)
+
+      begin
+        yield loader
+      ensure
+        loader.close
+      end
+    end
+
+    attr_reader :closed
+
+    def initialize(input_pdf_path)
+      begin
+        @doc = PDDocument.load(JavaFile.new(input_pdf_path))
+      rescue
+        raise Error, $!.message
+      end
+
+      @pages = @doc.pages.to_a
+
+      @closed = false
+    end
+
+    def load(entries)
+      root_outline =  PDDocumentOutline.new
+
+      @doc.document_catalog.document_outline = root_outline
+
+      entries.each do |entry|
+        set_outline(root_outline, entry)
+      end
+    end
+
+    def save(output_pdf_path)
+      @doc.save(output_pdf_path)
+    end
+
+    def close
+      unless @closed
+        @doc.close
+        @closed = true
+      end
+    end
+
+    private
+
+    def set_outline(parent_outline, entry)
+      page_number = entry.fetch('page')
+
+      max_page_number = @pages.size
+
+      unless (1..max_page_number).include?(page_number)
+        raise Error, "page number must be between 1 and #{max_page_number}: got #{page_number}"
+      end
+
+      page = @pages[page_number - 1]
+
+      dest = PDPageXYZDestination.new
+      dest.page = page
+      dest.top = page.bbox.upper_right_y
+      dest.left = 0
+      dest.zoom = -1
+
+      outline = PDOutlineItem.new
+      outline.destination = dest
+      outline.title = entry.fetch('title')
+
+      entry.fetch('children', []).each do |child|
+        set_outline(outline, child)
+      end
+
+      parent_outline.add_last(outline)
+    end
+  end
+end

data/lib/pdf_outline_editor/runner.rb

@@ -0,0 +1,56 @@
+require 'optparse'
+
+module PdfOutlineEditor
+  class Runner
+    class << self
+      _known_commands = {}
+
+      define_method :known_commands do
+        _known_commands
+      end
+
+      define_method :register_command do |key, klass|
+        _known_commands[key] = klass
+      end
+    end
+
+    def initialize(argv)
+      @argv = argv
+
+      @parser = nil
+
+      parser.order!(@argv)
+    end
+
+    def run
+      if @argv.empty?
+        raise Error, "missing a command: available commands are #{Runner.known_commands.keys.sort.join(', ')}"
+      end
+
+      command_name = @argv.shift
+
+      command_class = Runner.known_commands[command_name.to_sym]
+
+      if command_class
+        command = command_class.new(@argv)
+        command.run
+      else
+        raise Error, "unknown command: #{command_name}"
+      end
+    end
+
+    private
+
+    def parser
+      return @parser if @parser
+
+      @parser = OptionParser.new
+
+      @parser.banner = "Usage: #{File.basename($0)} <command> <args>"
+
+      @parser.version = VERSION
+
+      @parser
+    end
+  end
+end

data/lib/pdf_outline_editor/version.rb

@@ -0,0 +1,3 @@
+module PdfOutlineEditor
+  VERSION = "0.1.0"
+end

metadata

@@ -0,0 +1,98 @@
+--- !ruby/object:Gem::Specification
+name: pdf_outline_editor
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: java
+authors:
+- healthypackrat
+autorequire:
+bindir: exe
+cert_chain: []
+date: 2019-03-31 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.0'
+  name: bundler
+  prerelease: false
+  type: :development
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.0'
+- !ruby/object:Gem::Dependency
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '10.0'
+  name: rake
+  prerelease: false
+  type: :development
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '10.0'
+- !ruby/object:Gem::Dependency
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.0'
+  name: rspec
+  prerelease: false
+  type: :development
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.0'
+description:
+email:
+- healthypackrat@gmail.com
+executables:
+- pdf_outline_editor
+extensions: []
+extra_rdoc_files: []
+files:
+- README.md
+- exe/pdf_outline_editor
+- lib/pdf_outline_editor.rb
+- lib/pdf_outline_editor/commands/dump_command.rb
+- lib/pdf_outline_editor/commands/init_command.rb
+- lib/pdf_outline_editor/commands/load_command.rb
+- lib/pdf_outline_editor/dumper.rb
+- lib/pdf_outline_editor/loader.rb
+- lib/pdf_outline_editor/runner.rb
+- lib/pdf_outline_editor/version.rb
+- vendor/pdfbox/pdfbox-app-2.0.13.jar
+homepage: https://github.com/healthypackrat/pdf_outline_editor
+licenses:
+- Apache-2.0
+metadata: {}
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project:
+rubygems_version: 2.7.6
+signing_key:
+specification_version: 4
+summary: Set PDF outlines from a JSON/YAML definition file
+test_files: []