# -*- encoding: utf-8 -*-
#
#--
# This file is part of HexaPDF.
#
# HexaPDF - A Versatile PDF Creation and Manipulation Library For Ruby
# Copyright (C) 2014-2017 Thomas Leitner
#
# HexaPDF is free software: you can redistribute it and/or modify it
# under the terms of the GNU Affero General Public License version 3 as
# published by the Free Software Foundation with the addition of the
# following permission added to Section 15 as permitted in Section 7(a):
# FOR ANY PART OF THE COVERED WORK IN WHICH THE COPYRIGHT IS OWNED BY
# THOMAS LEITNER, THOMAS LEITNER DISCLAIMS THE WARRANTY OF NON
# INFRINGEMENT OF THIRD PARTY RIGHTS.
#
# HexaPDF is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public
# License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with HexaPDF. If not, see <http://www.gnu.org/licenses/>.
#
# The interactive user interfaces in modified source and object code
# versions of HexaPDF must display Appropriate Legal Notices, as required
# under Section 5 of the GNU Affero General Public License version 3.
#
# In accordance with Section 7(b) of the GNU Affero General Public
# License, a covered work must retain the producer line in every PDF that
# is created or manipulated using HexaPDF.
#++

require 'stringio'
require 'hexapdf/error'
require 'hexapdf/content'
require 'hexapdf/configuration'
require 'hexapdf/reference'
require 'hexapdf/object'
require 'hexapdf/stream'
require 'hexapdf/revisions'
require 'hexapdf/type'
require 'hexapdf/task'
require 'hexapdf/encryption'
require 'hexapdf/writer'
require 'hexapdf/importer'
require 'hexapdf/image_loader'
require 'hexapdf/font_loader'

# == HexaPDF API Documentation
#
# Here are some pointers to more in depth information:
#
# * For information about the command line application, see the HexaPDF::CLI module.
# * HexaPDF::Document provides information about how to work with a PDF file.
# * HexaPDF::Content::Canvas provides the canvas API for drawing/writing on a page or form XObject
module HexaPDF

  # Represents one PDF document.
  #
  # A PDF document consists of (indirect) objects, so the main job of this class is to provide
  # methods for working with these objects. However, since a PDF document may also be
  # incrementally updated and can therefore contain one or more revisions, there are also methods
  # to work with these revisions.
  #
  # Note: This class provides everything to work on PDF documents on a low-level basis. This means
  # that there are no convenience methods for higher PDF functionality whatsoever.
  class Document

    autoload(:Pages, 'hexapdf/document/pages')
    autoload(:Fonts, 'hexapdf/document/fonts')
    autoload(:Images, 'hexapdf/document/images')
    autoload(:Files, 'hexapdf/document/files')

    # :call-seq:
    #   Document.open(filename, **docargs)                   -> doc
    #   Document.open(filename, **docargs) {|doc| block}     -> obj
    #
    # Creates a new PDF Document object for the given file.
    #
    # Depending on whether a block is provided, the functionality is different:
    #
    # * If no block is provided, the whole file is instantly read into memory and the PDF Document
    #   created for it is returned.
    #
    # * If a block is provided, the file is opened and a PDF Document is created for it. The
    #   created document is passed as an argument to the block and when the block returns the
    #   associated file object is closed. The value of the block will be returned.
    #
    # The block version is useful, for example, when you are dealing with a large file and you
    # only need a small portion of it.
    #
    # The provided keyword arguments (except +io+) are passed on unchanged to Document.new.
    def self.open(filename, **kwargs)
      if block_given?
        File.open(filename, 'rb') do |file|
          yield(new(**kwargs, io: file))
        end
      else
        new(**kwargs, io: StringIO.new(File.binread(filename)))
      end
    end

    # The configuration for the document.
    attr_reader :config

    # The revisions of the document.
    attr_reader :revisions

    # Creates a new PDF document, either an empty one or one read from the provided +io+.
    #
    # When an IO object is provided and it contains an encrypted PDF file, it is automatically
    # decrypted behind the scenes. The +decryption_opts+ argument has to be set appropriately in
    # this case.
    #
    # Options:
    #
    # io:: If an IO object is provided, then this document can read PDF objects from this IO
    #      object, otherwise it can only contain created PDF objects.
    #
    # decryption_opts:: A hash with options for decrypting the PDF objects loaded from the IO.
    #
    # config:: A hash with configuration options that is deep-merged into the default
    #          configuration (see DefaultDocumentConfiguration), meaning that direct sub-hashes
    #          are merged instead of overwritten.
    def initialize(io: nil, decryption_opts: {}, config: {})
      @config = Configuration.with_defaults(config)
      @version = '1.2'

      @revisions = Revisions.from_io(self, io)
      if encrypted? && @config['document.auto_decrypt']
        @security_handler = Encryption::SecurityHandler.set_up_decryption(self, decryption_opts)
      else
        @security_handler = nil
      end

      @listeners = {}
    end

    # :call-seq:
    #   doc.object(ref)    -> obj or nil
    #   doc.object(oid)    -> obj or nil
    #
    # Returns the current version of the indirect object for the given exact reference or for the
    # given object number.
    #
    # For references to unknown objects, +nil+ is returned but free objects are represented by a
    # PDF Null object, not by +nil+!
    #
    # See: PDF1.7 s7.3.9
    def object(ref)
      i = @revisions.size - 1
      while i >= 0
        return @revisions[i].object(ref) if @revisions[i].object?(ref)
        i -= 1
      end
      nil
    end

    # Dereferences the given object.
    #
    # Return the object itself if it is not a reference, or the indirect object specified by the
    # reference.
    def deref(obj)
      obj.kind_of?(Reference) ? object(obj) : obj
    end

    # :call-seq:
    #   doc.object?(ref)    -> true or false
    #   doc.object?(oid)    -> true or false
    #
    # Returns +true+ if the the document contains an indirect object for the given exact reference
    # or for the given object number.
    #
    # Even though this method might return +true+ for some references, #object may return +nil+
    # because this method takes *all* revisions into account. Also see the discussion on #each for
    # more information.
    def object?(ref)
      @revisions.any? {|rev| rev.object?(ref)}
    end

    # :call-seq:
    #   doc.add(obj, revision: :current, **wrap_opts)     -> indirect_object
    #
    # Adds the object to the specified revision of the document and returns the wrapped indirect
    # object.
    #
    # The object can either be a native Ruby object (Hash, Array, Integer, ...) or a
    # HexaPDF::Object. If it is not the latter, #wrap is called with the object and the
    # additional keyword arguments.
    #
    # If the +revision+ option is +:current+, the current revision is used. Otherwise +revision+
    # should be a revision index.
    def add(obj, revision: :current, **wrap_opts)
      obj = wrap(obj, wrap_opts) unless obj.kind_of?(HexaPDF::Object)

      revision = (revision == :current ? @revisions.current : @revisions.revision(revision))
      if revision.nil?
        raise ArgumentError, "Invalid revision index specified"
      end

      if obj.document? && obj.document != self
        raise HexaPDF::Error, "Can't add object that is already attached to another document"
      end
      obj.document = self

      if obj.indirect? && (rev_obj = revision.object(obj.oid))
        if rev_obj.equal?(obj)
          return obj
        else
          raise HexaPDF::Error, "Can't add object because the specified revision already has " \
            "an object with object number #{obj.oid}"
        end
      end

      obj.oid = @revisions.map(&:next_free_oid).max unless obj.indirect?

      revision.add(obj)
    end

    # :call-seq:
    #   doc.delete(ref, revision: :all)
    #   doc.delete(oid, revision: :all)
    #
    # Deletes the indirect object specified by an exact reference or by an object number from the
    # document.
    #
    # Options:
    #
    # revision:: Specifies from which revisions the object should be deleted:
    #
    #            :all:: Delete the object from all revisions.
    #            :current:: Delete the object only from the current revision.
    #
    # mark_as_free:: If +true+, objects are only marked as free objects instead of being actually
    #                deleted.
    def delete(ref, revision: :all, mark_as_free: true)
      case revision
      when :current
        @revisions.current.delete(ref, mark_as_free: mark_as_free)
      when :all
        @revisions.each {|rev| rev.delete(ref, mark_as_free: mark_as_free)}
      else
        raise ArgumentError, "Unsupported option revision: #{revision}"
      end
    end

    # :call-seq:
    #   doc.import(obj)     -> imported_object
    #
    # Imports the given, with a different document associated PDF object and returns the imported
    # object.
    #
    # If the same argument is provided in multiple invocations, the import is done only once and
    # the previously imoprted object is returned.
    #
    # See: Importer
    def import(obj)
      if !obj.kind_of?(HexaPDF::Object) || !obj.document? || obj.document == self
        raise ArgumentError, "Importing only works for PDF objects associated " \
          "with another document"
      end
      HexaPDF::Importer.for(source: obj.document, destination: self).import(obj)
    end

    # Wraps the given object inside a HexaPDF::Object class which allows one to use
    # convenience functions to work with the object.
    #
    # The +obj+ argument can also be a HexaPDF::Object object so that it can be re-wrapped if
    # needed.
    #
    # The class of the returned object is always a subclass of HexaPDF::Object (or of
    # HexaPDF::Stream if a +stream+ is given). Which subclass is used, depends on the values
    # of the +type+ and +subtype+ options as well as on the 'object.type_map' and
    # 'object.subtype_map' global configuration options:
    #
    # * If *only* +type+ or +subtype+ is provided and a mapping is found, the resulting class is
    #   used.
    #
    # * If both +type+ and +subtype+ are provided and and a mapping for +subtype+ is found, the
    #   resulting class is used. If no mapping is found but there is a mapping for +type+, the
    #   mapped class is used.
    #
    # * If there is no valid class after the above steps, HexaPDF::Stream is used if a stream
    #   is given, HexaPDF::Dictionary is used if the given objecct is a hash or else
    #   HexaPDF::Object is used.
    #
    # Options:
    #
    # :type:: (Symbol or Class) The type of a PDF object that should be used for wrapping. This
    #         could be, for example, :Pages. If a class object is provided, it is used directly
    #         instead of the type detection system.
    #
    # :subtype:: (Symbol) The subtype of a PDF object which further qualifies a type. For
    #            example, image objects in PDF have a type of :XObject and a subtype of :Image.
    #
    # :oid:: (Integer) The object number that should be set on the wrapped object. Defaults to 0
    #        or the value of the given object's object number.
    #
    # :gen:: (Integer) The generation number that should be set on the wrapped object. Defaults to
    #        0 or the value of the given object's generation number.
    #
    # :stream:: (String or StreamData) The stream object which should be set on the wrapped
    #           object.
    def wrap(obj, type: nil, subtype: nil, oid: nil, gen: nil, stream: nil)
      data = if obj.kind_of?(HexaPDF::Object)
               obj.data
             else
               HexaPDF::PDFData.new(obj)
             end
      data.oid = oid if oid
      data.gen = gen if gen
      data.stream = stream if stream

      if type.kind_of?(Class)
        klass = type
      else
        default = if data.stream
                    HexaPDF::Stream
                  elsif data.value.kind_of?(Hash)
                    HexaPDF::Dictionary
                  else
                    HexaPDF::Object
                  end
        if data.value.kind_of?(Hash)
          type ||= deref(data.value[:Type])
          subtype ||= deref(data.value[:Subtype])
        end

        if subtype
          klass = GlobalConfiguration.constantize('object.subtype_map'.freeze, subtype)
        end
        if type && !klass
          klass = GlobalConfiguration.constantize('object.type_map'.freeze, type)
        end
        klass ||= default
      end

      klass.new(data, document: self)
    end

    # :call-seq:
    #   document.unwrap(obj)   -> unwrapped_obj
    #
    # Recursively unwraps the object to get native Ruby objects (i.e. Hash, Array, Integer, ...
    # instead of HexaPDF::Reference and HexaPDF::Object).
    def unwrap(object, seen = {})
      object = deref(object)
      object = object.data if object.kind_of?(HexaPDF::Object)
      if seen.key?(object)
        raise HexaPDF::Error, "Can't unwrap a recursive structure"
      end

      case object
      when Hash
        seen[object] = true
        object.each_with_object({}) {|(key, val), memo| memo[key] = unwrap(val, seen.dup)}
      when Array
        seen[object] = true
        object.map {|inner_o| unwrap(inner_o, seen.dup)}
      when HexaPDF::PDFData
        seen[object] = true
        unwrap(object.value, seen.dup)
      else
        object
      end
    end

    # :call-seq:
    #   doc.each(current: true) {|obj| block }        -> doc
    #   doc.each(current: true) {|obj, rev| block }   -> doc
    #   doc.each(current: true)                       -> Enumerator
    #
    # Calls the given block once for every object in the PDF document. The block may either accept
    # only the object or the object and the revision it is in.
    #
    # By default, only the current version of each object is returned which implies that each
    # object number is yielded exactly once. If the +current+ option is +false+, all stored
    # objects from newest to oldest are returned, not only the current version of each object.
    #
    # The +current+ option can make a difference because the document can contain multiple
    # revisions:
    #
    # * Multiple revisions may contain objects with the same object and generation numbers, e.g.
    #   two (different) objects with oid/gen [3,0].
    #
    # * Additionally, there may also be objects with the same object number but different
    #   generation numbers in different revisions, e.g. one object with oid/gen [3,0] and one with
    #   oid/gen [3,1].
    def each(current: true, &block)
      return to_enum(__method__, current: current) unless block_given?

      yield_rev = (block.arity == 2)
      oids = {}
      @revisions.reverse_each do |rev|
        rev.each do |obj|
          next if current && oids.include?(obj.oid)
          (yield_rev ? yield(obj, rev) : yield(obj))
          oids[obj.oid] = true
        end
      end
      self
    end

    # :call-seq:
    #    doc.register_listener(name, callable)             -> callable
    #    doc.register_listener(name) {|*args| block}       -> block
    #
    # Registers the given listener for the message +name+.
    def register_listener(name, callable = nil, &block)
      callable ||= block
      (@listeners[name] ||= []) << callable
      callable
    end

    # Dispatches the message +name+ with the given arguments to all registered listeners.
    def dispatch_message(name, *args)
      @listeners[name] && @listeners[name].each {|obj| obj.call(*args)}
    end

    # Returns the Pages object that provides convenience methods for working with pages.
    #
    # Also see: HexaPDF::Type::PageTreeNode
    def pages
      @pages ||= Pages.new(self)
    end

    # Returns the Images object that provides convenience methods for working with images.
    def images
      @images ||= Images.new(self)
    end

    # Returns the Files object that provides convenience methods for working with files.
    def files
      @files ||= Files.new(self)
    end

    # Returns the Fonts object that provides convenience methods for working with fonts.
    def fonts
      @fonts ||= Fonts.new(self)
    end

    # Executes the given task and returns its result.
    #
    # Tasks provide an extensible way for performing operations on a PDF document without
    # cluttering the Document interface.
    #
    # See Task for more information.
    def task(name, **opts, &block)
      task = GlobalConfiguration.constantize('task.map'.freeze, name) do
        raise HexaPDF::Error, "No task named '#{name}' is available"
      end
      task.call(self, **opts, &block)
    end

    # Returns the trailer dictionary for the document.
    def trailer
      @revisions.current.trailer
    end

    # Returns the document's catalog, the root of the object tree.
    def catalog
      trailer.catalog
    end

    # Returns the PDF document's version as string (e.g. '1.4').
    #
    # This method takes the file header version and the catalog's /Version key into account. If a
    # version has been set manually and the catalog's /Version key refers to a later version, the
    # later version is used.
    #
    # See: PDF1.7 s7.2.2
    def version
      catalog_version = (catalog[:Version] || '1.0'.freeze).to_s
      (@version < catalog_version ? catalog_version : @version)
    end

    # Sets the version of the PDF document. The argument must be a string in the format 'M.N'
    # where M is the major version and N the minor version (e.g. '1.4' or '2.0').
    def version=(value)
      raise ArgumentError, "PDF version must follow format M.N" unless value.to_s =~ /\A\d\.\d\z/
      @version = value.to_s
    end

    # Returns +true+ if the document is encrypted.
    def encrypted?
      !trailer[:Encrypt].nil?
    end

    # Encrypts the document.
    #
    # This is done by setting up a security handler for this purpose and populating the trailer's
    # Encrypt dictionary accordingly. The actual encryption, however, is only done when writing the
    # document.
    #
    # The security handler used for encrypting is selected via the +name+ argument. All other
    # arguments are passed on the security handler.
    #
    # If the document should not be encrypted, the +name+ argument has to be set to +nil+. This
    # removes the security handler and deletes the trailer's Encrypt dictionary.
    #
    # See: HexaPDF::Encryption::SecurityHandler#set_up_encryption and
    # HexaPDF::Encryption::StandardSecurityHandler::EncryptionOptions for possible encryption
    # options.
    def encrypt(name: :Standard, **options)
      if name.nil?
        trailer.delete(:Encrypt)
        @security_handler = nil
      else
        @security_handler = Encryption::SecurityHandler.set_up_encryption(self, name, **options)
      end
    end

    # Returns the security handler that is used for decrypting or encrypting the document, or +nil+
    # if none is set.
    #
    # * If the document was created by reading an existing file and the document was automatically
    #   decrypted, then this method returns the handler for decrypting.
    #
    # * Once the #encrypt method is called, the specified security handler for encrypting is
    #   returned.
    def security_handler
      @security_handler
    end

    # :call-seq:
    #   doc.validate(auto_correct: true)                               -> true or false
    #   doc.validate(auto_correct: true) {|msg, correctable| block }   -> true or false
    #
    # Validates all objects of the document, with optional auto-correction, and returns +true+ if
    # everything is fine.
    #
    # If a block is given, it is called on validation problems.
    #
    # See HexaPDF::Object#validate for more information.
    def validate(auto_correct: true, &block)
      result = trailer.validate(auto_correct: auto_correct, &block)
      each(current: false) do |obj|
        result &&= obj.validate(auto_correct: auto_correct, &block)
      end
      result
    end

    # :call-seq:
    #   doc.write(filename, validate: true, update_fields: true, optimize: false)
    #   doc.write(io, validate: true, update_fields: true, optimize: false)
    #
    # Writes the document to the given file (in case +io+ is a String) or IO stream.
    #
    # Before the document is written, it is validated using #validate and an error is raised if the
    # document is not valid. However, this step can be skipped if needed.
    #
    # Options:
    #
    # validate::
    #   Validates the document and raises an error if an uncorrectable problem is found.
    #
    # update_fields::
    #   Updates the /ID field in the trailer dictionary as well as the /ModDate field in the
    #   trailer's /Info dictionary so that it is clear that the document has been updated.
    #
    # optimize::
    #   Optimize the file size by using object and cross-reference streams. This will raise the PDF
    #   version to at least 1.5.
    def write(file_or_io, validate: true, update_fields: true, optimize: false)
      dispatch_message(:complete_objects)

      if update_fields
        trailer.update_id
        trailer.info[:ModDate] = Time.now
      end

      if validate
        self.validate(auto_correct: true) do |msg, correctable|
          next if correctable
          raise HexaPDF::Error, "Validation error: #{msg}"
        end
      end

      if optimize
        task(:optimize, object_streams: :generate)
        self.version = '1.5' if version < '1.5'
      end

      dispatch_message(:before_write)

      if file_or_io.kind_of?(String)
        File.open(file_or_io, 'w+') {|file| Writer.write(self, file)}
      else
        Writer.write(self, file_or_io)
      end
    end

  end

end