doc2text 0.3.2 → 0.4.3

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: doc2text
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.4.3
 platform: ruby
 authors:
-- Valentin Aitken
+- Valentin A.
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-10-23 00:00:00.000000000 Z
+date: 2021-01-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -16,56 +16,54 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.6'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.6.3
+        version: 1.11.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.6'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.6.3
+        version: 1.11.1
 - !ruby/object:Gem::Dependency
   name: rubyzip
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.1.6
+        version: 2.3.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.1.6
+        version: 2.3.0
 description: Parses odt to markdown
-email: bostko@gmail.com
-executables: []
+email: valentin@nalisbg.com
+executables:
+- doc2text
 extensions: []
 extra_rdoc_files: []
 files:
+- bin/doc2text
 - lib/doc2text.rb
-- lib/doc2text/content.rb
+- lib/doc2text/docx/docx.rb
+- lib/doc2text/docx/docx_xml_namespaces.rb
+- lib/doc2text/docx/markdown_docx_parser.rb
 - lib/doc2text/errors.rb
-- lib/doc2text/markdown_odt_parser.rb
-- lib/doc2text/odt.rb
-- lib/doc2text/odt_xml_namespaces.rb
-- lib/doc2text/odt_xml_node.rb
+- lib/doc2text/generic_xml_nodes.rb
+- lib/doc2text/odt/markdown_odt_parser.rb
+- lib/doc2text/odt/odt.rb
+- lib/doc2text/odt/odt_xml_namespaces.rb
+- lib/doc2text/pptx/markdown_pptx_parser.rb
+- lib/doc2text/pptx/pptx.rb
+- lib/doc2text/pptx/pptx_xml_namespaces.rb
+- lib/doc2text/resolution.rb
+- lib/doc2text/styles_parser.rb
+- lib/doc2text/xml_based_document_file.rb
 homepage: http://doc2text.com
 licenses:
-- GPL
+- Apache-2.0
 metadata: {}
 post_install_message: 
 rdoc_options: []
@@ -82,8 +80,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.2.2
+rubygems_version: 3.1.2
 signing_key: 
 specification_version: 4
 summary: Translates odt to markdown

data/lib/doc2text/content.rb

@@ -1,25 +0,0 @@
-module Doc2Text
-  module Odt
-    module Content
-      class Document < ::Nokogiri::XML::SAX::Document
-        def initialize(markdown_odt_parser)
-          @markdown_odt_parser = markdown_odt_parser
-        end
-
-        def start_element_namespace(name ,attrs = [], prefix = nil, uri = nil, ns = [])
-          @markdown_odt_parser.new_node prefix, name, attrs
-        end
-
-        def end_element_namespace(name, prefix = nil, uri = nil)
-          @markdown_odt_parser.close_node prefix, name
-        end
-
-        def characters(string)
-          unless string.strip.empty?
-            @markdown_odt_parser.text string
-          end
-        end
-      end
-    end
-  end
-end

data/lib/doc2text/markdown_odt_parser.rb

@@ -1,72 +0,0 @@
-require 'logger'
-
-module Doc2Text
-  module Markdown
-    class OdtParser
-      def initialize(output)
-        @output = output
-        @automatic_styles = {}
-      end
-
-      def new_node(prefix, name, attrs)
-        unless @xml_root
-          @xml_root = @current_node = Odt::XmlNodes::Node.create_node prefix, name, nil, attrs, self
-        else
-          new_node = Odt::XmlNodes::Node.create_node prefix, name, @current_node, attrs, self
-          @current_node.children << new_node
-          @current_node = new_node
-        end
-      end
-
-      def close_node(prefix, name)
-        # if Odt::XmlNodes::Node.create_node(prefix, name, nil, [], self).eql? @current_node
-          if @current_node.parent and @current_node.parent.office_text?
-            @output << @current_node.expand
-            @current_node.delete
-          end
-          @current_node = @current_node.parent
-        # else
-        #   # TODO remove this redundant(tree build algorithm) checks
-        #   raise Doc2Text::XmlError, "!Close node child #{prefix} #{name} IS NOT correct, CURRENT_ELEM #{@current_node}"
-        # end
-      end
-
-      def text(string)
-        plain_text = Odt::XmlNodes::PlainText.new(string)
-        @current_node.children << plain_text
-      end
-
-      def close
-        @output.close
-      end
-
-      def print_tree(node)
-        puts node
-        node.children.each do |child|
-          print_tree child
-        end
-      end
-
-      # Select nodes xpath style
-      # - supports selecting from the root node
-      def xpath(string)
-        if /^(\/[\w:\-]+)+$/ =~ string
-          path = string.scan /[\w:\-]+/
-          seek_nodes = [@xml_root]
-          path.each_with_index do |xml_name, index|
-            seek_nodes.select! { |node| node.xml_name == xml_name }
-            seek_nodes = seek_nodes.map(&:children).flatten unless index == path.length - 1
-            break if seek_nodes.empty?
-          end
-          seek_nodes
-        else
-          raise Doc2Text::XmlError, 'it does not support this xpath syntax'
-        end
-      end
-
-      def logger
-        @logger ||= Logger.new(STDOUT)
-      end
-    end
-  end
-end

data/lib/doc2text/odt.rb

@@ -1,73 +0,0 @@
-require 'zip'
-
-module Doc2Text
-  module Odt
-    class Document
-      EXTRACT_EXTENSION = 'unpacked_odt'
-
-      def self.parse_and_save(input, output_filename)
-        odt = new input
-        begin
-          odt.unpack
-          output = File.open output_filename, 'w'
-          markdown = Markdown::OdtParser.new output
-          begin
-            odt.parse markdown
-          ensure
-            markdown.close
-          end
-        ensure
-          odt.clean
-        end
-      end
-
-      def parse(markdown)
-        content = ::Doc2Text::Odt::Content::Document.new markdown
-        parser = Nokogiri::XML::SAX::Parser.new(content) # { |config| config.strict}
-        parser.parse open 'content.xml'
-      end
-
-      def initialize(document_path)
-        @document_path = document_path
-      end
-
-      def unpack
-        Zip::File.open(@document_path) {
-            |zip_file|
-          Dir.mkdir(extract_path)
-          zip_file.each do |entry|
-            zipped_file_extract_path = File.join extract_path, entry.name
-            FileUtils.mkdir_p File.dirname(zipped_file_extract_path)
-            zip_file.extract entry, zipped_file_extract_path
-          end
-        }
-      end
-
-      def clean
-        if [extract_path, File.join(extract_path, 'content.xml'), File.join(extract_path, 'mimetype')].all? { |file| File.exist?(file) }
-          FileUtils.rm_r extract_path
-        end
-      end
-
-      # Open file from the current odt
-      def open(filename)
-        File.open File.join(extract_path, filename), 'r'
-      end
-
-      # Parse xml file from the current odt
-      def xml_file(filename, rood_node_name)
-        Nokogiri::XML::Document.parse(open(filename)) { |config| config.strict }
-        root_node = doc.root
-        if root_node.name != rood_node_name or root_node.namespace.prefix != 'office'
-          raise XmlError, 'Document does not have correct root element'
-        else
-          open(filename)
-        end
-      end
-
-      def extract_path
-        File.join File.dirname(@document_path), ".#{File.basename(@document_path)}_#{EXTRACT_EXTENSION}"
-      end
-    end
-  end
-end

data/lib/doc2text/odt_xml_namespaces.rb

@@ -1,215 +0,0 @@
-module Doc2Text
-  module Odt
-    module XmlNodes
-      class PlainText
-        include Node
-
-        attr_accessor :text
-
-        alias_method :expand, :text
-
-        def initialize(text)
-          @text = text
-        end
-      end
-
-      class Generic
-        include Node
-      end
-
-      #
-      # These are the namespaces available in the open document format
-      # http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os.html
-      #
-      module Office
-        class AutomaticStyles
-          include Node
-        end
-
-        class DocumentContent
-          include Node
-        end
-
-        class Text
-          include Node
-
-          def office_text?
-            true
-          end
-        end
-      end
-
-      module Animation; end
-      module Chart; end
-      module Config; end
-      module Database; end
-      module Dr3d; end
-      module Drawing; end
-      module Form; end
-      module Manifest; end
-      module Meta; end
-      module DataStyle; end
-      module Presentation; end
-      module Script; end
-      module Table
-        class TableRow
-          include Node
-
-          def expand
-            header_delimiter = parent.children.count >= 2 && parent.children[1] == self ? "\n|---|---|" : ''
-            result = "\n#{@children.map(&:expand).join.strip.gsub "\n", ''} |#{header_delimiter}"
-            delete
-            result
-          end
-        end
-
-        class TableCell
-          include Node
-
-          def open
-            ' | '
-          end
-        end
-      end
-      module Style
-        class Style
-          include Node
-        end
-
-        class TextProperties
-          include Node
-        end
-      end
-      module XslFoCompatible; end
-      module SvgCompatible; end
-      module SmilCompatible; end
-      module Of; end
-
-      module Text
-        def initialize(parent = nil, attrs = [], prefix = nil, name = nil, markdown_odt_parser = nil)
-          super parent, attrs, prefix, name
-          @markdown_odt_parser = markdown_odt_parser
-          style_index = attrs.index { |attr| attr.prefix == 'text' && attr.localname == 'style-name' }
-          @enclosing_style = []
-          if style_index and fetch_style?
-            elem_style = find_style attrs[style_index].value
-            fetch_style elem_style
-          end
-        end
-
-        def fetch_style?
-          true
-        end
-
-        def fetch_style(elem_style)
-          if elem_style
-            elem_style.children.select { |style_property| style_property.xml_name == 'style:text-properties' }.each { |text_property|
-              text_property.attrs.each { |attr|
-                if attr.prefix == 'style'
-                  if attr.localname == 'font-style-complex' && attr.value == 'italic'
-                    @enclosing_style << '_'
-                  elsif attr.localname == 'font-weight-complex' && attr.value == 'bold'
-                    @enclosing_style << '**'
-                  end
-                end
-              }
-            }
-          end
-        end
-
-        def find_style(style_name)
-          styles = @markdown_odt_parser.xpath '/office:document-content/office:automatic-styles/style:style'
-          styles.find { |style| style.attrs.index { |attr| attr.prefix == 'style' && attr.localname == 'family' } &&
-              style.attrs.index { |attr| attr.prefix == 'style' && attr.localname == 'name' && attr.value == style_name } }
-        end
-
-        # http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part1.html#__RefHeading__1419256_253892949
-        class P
-          include Node
-          include Text
-
-          def initialize(parent = nil, attrs = [], prefix = nil, name = nil, markdown_odt_parser = nil)
-            super parent, attrs, prefix, name, markdown_odt_parser
-          end
-
-          def self.style_family
-            'paragraph'
-          end
-
-          def open
-            "\n#{@enclosing_style.join}"
-          end
-
-          def close
-            "#{@enclosing_style.reverse.join}\n"
-          end
-        end
-
-        class LineBreak
-          include Node
-
-          def open
-            '<br/>'
-          end
-        end
-
-        # http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part1.html#__RefHeading__1419264_253892949
-        class Span
-          include Node
-          include Text
-
-          def self.style_family
-            'text'
-          end
-
-          def open
-            @enclosing_style.join
-          end
-
-          def close
-            @enclosing_style.reverse.join
-          end
-        end
-
-        # http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part1.html#__RefHeading__1415154_253892949
-        class ListItem
-          include Node
-          include Text
-
-          def expand
-            result = "* #{@children.map(&:expand).join.strip.gsub /\n{2,}/, "\n"}\n"
-            delete
-            result.clone
-          end
-
-          def fetch_style?
-            false
-          end
-        end
-
-        class List
-          include Node
-          include Text
-
-          def open
-            "\n"
-          end
-
-          def fetch_style(elem_style)
-            if elem_style
-              elem_style.children.select { |style_property| style_property.xml_name == 'style:text-properties' }.each { |text_property|
-                text_property.attrs.each { |attr|
-                  if attr.prefix == 'style'
-                    if attr.localname == 'list-level-style-number' && attr.value == 'Numbering_20_Symbols'
-                      @enclosing_style << '_'
-                    end
-                  end
-                }
-              }
-            end
-          end
-        end
-      end
-    end
-  end
-end