pdfh 0.2.1 → 3.0.1

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pdfh
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 3.0.1
 platform: ruby
 authors:
 - Isaias Piña
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-01-06 00:00:00.000000000 Z
+date: 2024-01-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: colorize
@@ -37,7 +37,7 @@ files:
 - ".rspec"
 - ".rubocop.yml"
 - ".rubocop_todo.yml"
-- ".ruby-version"
+- ".tool-versions"
 - CHANGELOG.md
 - CODE_OF_CONDUCT.md
 - Gemfile
@@ -51,15 +51,18 @@ files:
 - exe/pdfh
 - lib/ext/string.rb
 - lib/pdfh.rb
-- lib/pdfh/document.rb
-- lib/pdfh/document_period.rb
-- lib/pdfh/document_processor.rb
-- lib/pdfh/document_type.rb
-- lib/pdfh/month.rb
-- lib/pdfh/opt_parser.rb
-- lib/pdfh/pdf_handler.rb
-- lib/pdfh/settings.rb
+- lib/pdfh/main.rb
+- lib/pdfh/models/document.rb
+- lib/pdfh/models/document_period.rb
+- lib/pdfh/models/document_type.rb
+- lib/pdfh/models/settings.rb
 - lib/pdfh/settings_template.rb
+- lib/pdfh/utils/console.rb
+- lib/pdfh/utils/month.rb
+- lib/pdfh/utils/opt_parser.rb
+- lib/pdfh/utils/options.rb
+- lib/pdfh/utils/pdf_file_handler.rb
+- lib/pdfh/utils/settings_builder.rb
 - lib/pdfh/version.rb
 - pdfh.gemspec
 homepage: https://github.com/iax7/pdfh
@@ -79,14 +82,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 2.5.0
+      version: 3.0.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.3
+rubygems_version: 3.5.3
 signing_key:
 specification_version: 4
 summary: Organize PDF files

data/.ruby-version

@@ -1 +0,0 @@
-ruby-3.1.0

data/lib/pdfh/document_processor.rb

@@ -1,164 +0,0 @@
-# frozen_string_literal: true
-
-require "fileutils"
-
-module Pdfh
-  # Main functionality. This class is intended to manage the pdf documents
-  class DocumentProcessor
-    # @return [self]
-    def initialize
-      @options = Pdfh.parse_argv
-      Pdfh.verbose = options[:verbose]
-      Pdfh.dry = options[:dry]
-      Pdfh.verbose_print(options)
-      Pdfh.mode = options.key?(:type) ? :file : :directory
-    end
-
-    # @return [void]
-    def start
-      @settings = Settings.new(Pdfh.search_config_file)
-      puts "Destination path: #{@settings.base_path.colorize(:light_blue)}" if Pdfh.verbose?
-
-      Pdfh.file_mode? ? process_files : process_lookup_dirs
-    rescue SettingsIOError => e
-      Pdfh.error_print(e.message, exit_app: false)
-      Pdfh.create_settings_file
-      exit(1)
-    rescue StandardError => e
-      Pdfh.error_print e.message
-    end
-
-    private
-
-    attr_reader :options
-
-    # @param [String] file_name
-    # @return [DocumentType]
-    def match_doc_type(file_name)
-      @settings.document_types.each do |type|
-        match = type.re_file.match(file_name)
-        return type if match
-      end
-      nil
-    end
-
-    # @return [DocumentType]
-    def doc_type_by_id(id)
-      @settings.document_types.find { |t| t.gid == id }
-    end
-
-    # @return [void]
-    def process_files
-      type_id = options[:type]
-      raise ArgumentError, "No files provided to process #{type_id.inspect} type." unless options[:files]
-
-      type = doc_type_by_id(type_id)
-      Pdfh.error_print "Type #{type_id.inspect} was not found." if type.nil?
-      puts
-      options[:files].each do |file|
-        unless File.exist?(file)
-          Pdfh.warn_print "File #{file.inspect} does not exist."
-          next
-        end
-        unless File.extname(file) == ".pdf"
-          Pdfh.warn_print "File #{file.inspect} is not a pdf."
-          next
-        end
-        process_document(file, type)
-      end
-    end
-
-    # @return [void]
-    def process_lookup_dirs
-      @settings.lookup_dirs.each do |work_directory|
-        process_directory(work_directory)
-      end
-    end
-
-    # @param [String] work_directory
-    # @return [Enumerator]
-    def process_directory(work_directory)
-      Pdfh.headline(work_directory)
-      processed_count = 0
-      ignored_files = []
-      files = Dir["#{work_directory}/*.pdf"]
-      files.each do |pdf_file|
-        type = match_doc_type(pdf_file)
-        if type
-          processed_count += 1
-          process_document(pdf_file, type)
-        else
-          ignored_files << basename_without_ext(pdf_file)
-        end
-      end
-      puts "  (No files processed)".colorize(:light_black) if processed_count.zero?
-      return unless Pdfh.verbose?
-
-      puts "\n  No document type found for these PDF files:" if ignored_files.any?
-      ignored_files.each.with_index(1) { |file, index| Pdfh.ident_print index, file, color: :magenta }
-    end
-
-    ##
-    # Generate document, and process actions
-    # @param [String] file
-    # @param [DocumentType] type
-    # @return [void]
-    def process_document(file, type)
-      base = File.basename(file)
-      puts "Working on #{base.colorize(:light_green)}"
-      pad = 12
-      Pdfh.ident_print "Type", type.name, color: :light_blue, width: pad
-      doc = Document.new(file, type)
-      Pdfh.ident_print "Sub-Type", doc.sub_type, color: :light_blue, width: pad
-      Pdfh.ident_print "Period", doc.period.to_s, color: :light_blue, width: pad
-      Pdfh.ident_print "New Name", doc.new_name, color: :light_blue, width: pad
-      Pdfh.ident_print "Store Path", doc.store_path, color: :light_blue, width: pad
-      Pdfh.ident_print "Other files", doc.companion_files(join: true), color: :light_blue, width: pad
-      Pdfh.ident_print "Print CMD", doc.print_cmd, color: :light_blue, width: pad
-      Pdfh.ident_print "Processed?", "No (in Dry mode)", color: :red, width: pad if Pdfh.dry?
-      write_pdf(doc)
-    rescue StandardError => e
-      Pdfh.ident_print "Doc Error", e.message, color: :red, width: pad
-    end
-
-    def write_pdf(document)
-      base_path = @settings.base_path
-      full_path = File.join(base_path, document.store_path, document.new_name)
-      dir_path = File.join(base_path, document.store_path)
-
-      FileUtils.mkdir_p(dir_path) unless File.exist?(dir_path)
-
-      document.pdf_doc.write_new_pdf(dir_path, full_path)
-      make_document_backup(document)
-      copy_companion_files(dir_path, document)
-    end
-
-    # Create a backup of original document
-    def make_document_backup(document)
-      Pdfh.verbose_print "~~~~~~~~~~~~~~~~~~ Creating PDF backup"
-      Dir.chdir(document.home_dir) do
-        Pdfh.verbose_print "  Working on: #{document.home_dir.inspect} directory"
-        Pdfh.verbose_print "    mv #{document.file_name.inspect} -> #{document.backup_name.inspect}"
-        File.rename(document.file_name, document.backup_name) unless Pdfh.dry?
-      end
-    end
-
-    def copy_companion_files(destination, document)
-      Pdfh.verbose_print "~~~~~~~~~~~~~~~~~~ Writing Companion files"
-      document.companion_files.each do |file|
-        Pdfh.verbose_print "  Working on #{file.inspect}..."
-        src_name = File.join(document.home_dir, file)
-        src_ext = File.extname(file)
-        dest_name = File.basename(document.new_name, ".pdf")
-        dest_full = File.join(destination, "#{dest_name}#{src_ext}")
-        Pdfh.verbose_print "    cp #{src_name} --> #{dest_full}"
-        FileUtils.cp(src_name, dest_full) unless Pdfh.dry?
-      end
-    end
-
-    # @return [String]
-    def basename_without_ext(file)
-      File.basename(file, File.extname(file))
-    end
-  end
-end

data/lib/pdfh/opt_parser.rb

@@ -1,41 +0,0 @@
-# frozen_string_literal: true
-
-require "optparse"
-
-module Pdfh
-  OPT_PARSER = OptionParser.new do |opts|
-    opts.default_argv
-    # Process ARGV
-    opts.banner = "Usage: #{opts.program_name} [options] [file1 ...]"
-    opts.separator ""
-    opts.separator "Specific options:"
-
-    opts.on("-tID", "--type=ID", "Document type id (requires a trailing file list)")
-    opts.on_tail("-T", "--list-types", "List document types in configuration") do
-      settings = Settings.new(Pdfh.search_config_file)
-      ident = 4
-      max_width = settings.document_types.map { |t| t.gid.size }.max
-      puts "#{" " * ident}#{"ID".ljust(max_width)}  Type Name"
-      puts "#{" " * ident}#{"-" * max_width}  -----------------------"
-      settings.document_types.each do |type|
-        puts "#{" " * ident}#{type.gid.ljust(max_width)}  #{type.name.inspect}"
-      end
-      exit
-    rescue SettingsIOError => e
-      Pdfh.error_print(e.message, exit_app: false)
-      Pdfh.create_settings_file
-      exit(1)
-    end
-    opts.on_tail("-V", "--version", "Show version") do
-      puts "#{opts.program_name} v#{Pdfh::VERSION}"
-      exit
-    end
-    opts.on_tail("-h", "--help", "help (this dialog)") do
-      puts opts
-      exit
-    end
-
-    opts.on("-v", "--verbose", "Show more output. Useful for debug")
-    opts.on("-d", "--dry", "Dry run, does not write new pdf")
-  end
-end

data/lib/pdfh/pdf_handler.rb

@@ -1,55 +0,0 @@
-# frozen_string_literal: true
-
-module Pdfh
-  # Handles the Pdf document text extraction and password removal
-  # TODO: Replace command utils with this gem
-  #  require 'pdf-reader'
-  #
-  #  reader = PDF::Reader.new(temp)
-  #  reader.pages.each do |page|
-  #      @text << page.text
-  #  end
-  class PdfHandler
-    attr_reader :file
-
-    # @return [self]
-    def initialize(file, password)
-      @file = file
-      @password_option = password ? "--password=#{password.inspect} " : ""
-    end
-
-    ##
-    # Gets the text from the pdf in order to execute
-    # the regular expresion matches
-    # @return [String]
-    def extract_text
-      temp = `mktemp`.chomp
-      Pdfh.verbose_print "~~~~~~~~~~~~~~~~~~ Extract PDF text"
-      Pdfh.verbose_print "  --> #{temp.inspect} temporal file assigned."
-
-      cmd = %(qpdf #{@password_option}--decrypt --stream-data=uncompress #{@file.inspect} #{temp.inspect})
-      Pdfh.verbose_print "  DeCrypt Command: #{cmd}"
-      _result = `#{cmd}`
-
-      cmd2 = %(pdftotext -enc UTF-8 #{temp.inspect} -)
-      Pdfh.verbose_print "  Extract Command: #{cmd2}"
-      text = `#{cmd2}`
-      Pdfh.verbose_print "  Text: #{text.inspect}"
-      text
-    end
-
-    # @return [void]
-    def write_new_pdf(dir_path, full_path)
-      Pdfh.verbose_print "~~~~~~~~~~~~~~~~~~ Writing PDFs"
-      raise IOError, "Path #{dir_path} not found." unless Dir.exist?(dir_path)
-
-      cmd = %(qpdf #{@password_option}--decrypt #{@file.inspect} #{full_path.inspect})
-      Pdfh.verbose_print "  Write PDF Command: #{cmd}"
-
-      return if Pdfh.dry?
-
-      _result = `#{cmd}`
-      raise IOError, "New PDF file #{full_path.inspect} was not created." unless File.file?(full_path)
-    end
-  end
-end