reverse_adoc 1.0.1 → 2.0.0

data/lib/reverse_adoc/converters/sup.rb

@@ -1,16 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Sup < Base
-      def to_coradoc(node, state = {})
-        content = treat_children(node, state)
-        Coradoc::Element::Inline::Superscript.new(content)
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-    end
-
-    register :sup, Sup.new
-  end
-end

data/lib/reverse_adoc/converters/table.rb

@@ -1,69 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Table < Base
-      def to_coradoc(node, state = {})
-        id = node["id"]
-        title = extract_title(node)
-        attrs = style(node)
-        content = treat_children_coradoc(node, state)
-        Coradoc::Element::Table.new(title, content, { id: id, attrs: attrs })
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-
-      def extract_title(node)
-        title = node.at("./caption")
-        return "" if title.nil?
-
-        treat_children(title, {})
-      end
-
-      def frame(node)
-        case node["frame"]
-        when "void"
-          "none"
-        when "hsides"
-          "topbot"
-        when "vsides"
-          "sides"
-        when "box", "border"
-          "all"
-        end
-      end
-
-      def rules(node)
-        case node["rules"]
-        when "all"
-          "all"
-        when "rows"
-          "rows"
-        when "cols"
-          "cols"
-        when "none"
-          "none"
-        end
-      end
-
-      def style(node)
-        attrs = Coradoc::Element::AttributeList.new
-        # Width is disabled on tables for now. (From #88)
-        # attrs.add_named("width", node["width"]) if node["width"]
-
-        frame_attr = frame(node)
-        attrs.add_named("frame", frame_attr) if frame_attr
-
-        rules_attr = rules(node)
-        attrs.add_named("rules", rules_attr) if rules_attr
-
-        # This line should be removed.
-        return "" if attrs.empty?
-
-        attrs
-      end
-    end
-
-    register :table, Table.new
-  end
-end

data/lib/reverse_adoc/converters/td.rb

@@ -1,83 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Td < Base
-      def to_coradoc(node, state = {})
-        # convert(node, state)
-        id = node["id"]
-        colrowattr = colrow(node["colspan"], node["rowspan"])
-        alignattr = alignstyle(node)
-        style = cellstyle(node)
-
-        singlepara = node.elements.size == 1 && node.elements.first.name == "p"
-        state[:tdsinglepara] = singlepara if singlepara
-
-        adoccell = node.at(".//ul | .//ol | .//pre | .//blockquote | .//br | .//hr") ||
-          node.at(".//p") && !singlepara
-
-        style = "a" if adoccell
-        delim = adoccell ? "\n" : " "
-        content = treat_children_coradoc(node, state)
-        options = {}.tap do |hash|
-          hash[:id] = id
-          hash[:colrowattr] = colrowattr
-          hash[:alignattr] = alignattr
-          hash[:style] = style
-          hash[:content] = content
-          hash[:delim] = delim
-        end
-
-        Coradoc::Element::Table::Cell.new(options)
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-
-      def cellstyle(_node)
-        ""
-      end
-
-      def colrow(colspan, rowspan)
-        if colspan && rowspan && colspan != "1" && rowspan != "1"
-          "#{colspan}.#{rowspan}+"
-        elsif colspan && colspan != "1"
-          "#{colspan}+"
-        elsif rowspan && rowspan != "1"
-          ".#{rowspan}+"
-        else
-          ""
-        end
-      end
-
-      def alignstyle(node)
-        align = node["align"]
-        valign = node["valign"]
-        a = case align
-            when "left" then "<"
-            when "center" then "^"
-            when "right" then ">"
-            else
-              ""
-            end
-        v = case valign
-            when "top" then "<"
-            when "middle" then "^"
-            when "bottom" then ">"
-            else
-              ""
-            end
-        if align && valign
-          "#{a}.#{v}"
-        elsif align
-          a
-        elsif valign
-          ".#{v}"
-        else
-          ""
-        end
-      end
-    end
-
-    register :td, Td.new
-  end
-end

data/lib/reverse_adoc/converters/text.rb

@@ -1,65 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Text < Base
-      def to_coradoc(node, state = {})
-        return treat_empty(node, state) if node.text.strip.empty?
-
-        Coradoc::Element::TextElement.new(treat_text(node))
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-
-      private
-
-      def treat_empty(node, state)
-        parent = node.parent.name.to_sym
-        if %i[ol ul].include?(parent) # Otherwise the identation is broken
-          ""
-        elsif state[:tdsinglepara]
-          ""
-        elsif node.text == " " # Regular whitespace text node
-          " "
-        else
-          ""
-        end
-      end
-
-      def treat_text(node)
-        text = node.text
-        text = preserve_nbsp(text)
-        text = remove_border_newlines(text)
-        text = remove_inner_newlines(text)
-        text = escape_keychars(text)
-
-        text = preserve_keychars_within_backticks(text)
-        escape_links(text)
-      end
-
-      def preserve_nbsp(text)
-        text.gsub(/\u00A0/, "&nbsp;")
-      end
-
-      def escape_links(text)
-        text.gsub(/<<([^>]*)>>/, "\\<<\\1>>")
-      end
-
-      def remove_border_newlines(text)
-        text.gsub(/\A\n+/, "").gsub(/\n+\z/, "")
-      end
-
-      def remove_inner_newlines(text)
-        text.tr("\n\t", " ").squeeze(" ")
-      end
-
-      def preserve_keychars_within_backticks(text)
-        text.gsub(/`.*?`/) do |match|
-          match.gsub('\_', "_").gsub('\*', "*")
-        end
-      end
-    end
-
-    register :text, Text.new
-  end
-end

data/lib/reverse_adoc/converters/th.rb

@@ -1,14 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Th < Td
-      def cellstyle(node)
-        # this is the header row
-        return "" if node.parent.previous_element.nil?
-
-        "h"
-      end
-    end
-
-    register :th, Th.new
-  end
-end

data/lib/reverse_adoc/converters/tr.rb

@@ -1,22 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Tr < Base
-      def to_coradoc(node, state = {})
-        content = treat_children_coradoc(node, state)
-        header = table_header_row?(node)
-        Coradoc::Element::Table::Row.new(content, header)
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-
-      def table_header_row?(node)
-        # node.element_children.all? {|child| child.name.to_sym == :th}
-        node.previous_element.nil?
-      end
-    end
-
-    register :tr, Tr.new
-  end
-end

data/lib/reverse_adoc/converters/video.rb

@@ -1,29 +0,0 @@
-module ReverseAdoc
-  module Converters
-    class Video < Base
-      def to_coradoc(node, _state = {})
-        src = node["src"]
-        id = node["id"]
-        title = extract_title(node)
-        attributes = Coradoc::Element::AttributeList.new
-        options = options(node)
-        attributes.add_named("options", options) if options.any?
-        Coradoc::Element::Video.new(title, id: id, src: src,
-                                           attributes: attributes)
-      end
-
-      def convert(node, state = {})
-        Coradoc::Generator.gen_adoc(to_coradoc(node, state))
-      end
-
-      def options(node)
-        autoplay = node["autoplay"]
-        loop_attr = node["loop"]
-        controls = node["controls"]
-        [autoplay, loop_attr, controls].compact
-      end
-    end
-
-    register :video, Video.new
-  end
-end

data/lib/reverse_adoc/converters.rb

@@ -1,32 +0,0 @@
-module ReverseAdoc
-  module Converters
-    def self.register(tag_name, converter)
-      @@converters ||= {}
-      @@converters[tag_name.to_sym] = converter
-    end
-
-    def self.unregister(tag_name)
-      @@converters.delete(tag_name.to_sym)
-    end
-
-    def self.lookup(tag_name)
-      @@converters[tag_name.to_sym] or default_converter(tag_name)
-    end
-
-    def self.default_converter(tag_name)
-      case ReverseAdoc.config.unknown_tags.to_sym
-      when :pass_through
-        ReverseAdoc::Converters::PassThrough.new
-      when :drop
-        ReverseAdoc::Converters::Drop.new
-      when :bypass
-        ReverseAdoc::Converters::Bypass.new
-      when :raise
-        raise UnknownTagError, "unknown tag: #{tag_name}"
-      else
-        raise InvalidConfigurationError,
-              "unknown value #{ReverseAdoc.config.unknown_tags.inspect} for ReverseAdoc.config.unknown_tags"
-      end
-    end
-  end
-end

data/lib/reverse_adoc/errors.rb

@@ -1,10 +0,0 @@
-module ReverseAdoc
-  class Error < StandardError
-  end
-
-  class UnknownTagError < Error
-  end
-
-  class InvalidConfigurationError < Error
-  end
-end

data/lib/reverse_adoc/html_converter.rb

@@ -1,61 +0,0 @@
-# frozen_string_literal: true
-
-require_relative "converters/a"
-require_relative "converters/aside"
-require_relative "converters/audio"
-require_relative "converters/blockquote"
-require_relative "converters/br"
-require_relative "converters/bypass"
-require_relative "converters/code"
-require_relative "converters/div"
-require_relative "converters/drop"
-require_relative "converters/em"
-require_relative "converters/figure"
-require_relative "converters/h"
-require_relative "converters/head"
-require_relative "converters/hr"
-require_relative "converters/ignore"
-require_relative "converters/img"
-require_relative "converters/mark"
-require_relative "converters/li"
-require_relative "converters/ol"
-require_relative "converters/p"
-require_relative "converters/pass_through"
-require_relative "converters/pre"
-require_relative "converters/q"
-require_relative "converters/strong"
-require_relative "converters/sup"
-require_relative "converters/sub"
-require_relative "converters/table"
-require_relative "converters/td"
-require_relative "converters/th"
-require_relative "converters/text"
-require_relative "converters/tr"
-require_relative "converters/video"
-require_relative "converters/math"
-
-module ReverseAdoc
-  class HtmlConverter
-    def self.to_coradoc(input, options = {})
-      root = case input
-             when String
-               Nokogiri::HTML(input).root
-             when Nokogiri::XML::Document
-               input.root
-             when Nokogiri::XML::Node
-               input
-             end
-
-      return "" unless root
-
-      ReverseAdoc.config.with(options) do
-        ReverseAdoc::Converters.lookup(root.name).to_coradoc(root)
-      end
-    end
-
-    def self.convert(input, options = {})
-      result = Coradoc::Generator.gen_adoc(to_coradoc(input, options))
-      ReverseAdoc.cleaner.tidy(result)
-    end
-  end
-end

data/lib/reverse_adoc/version.rb

@@ -1,3 +0,0 @@
-module ReverseAdoc
-  VERSION = "1.0.1".freeze
-end

data/lib/reverse_adoc.rb

@@ -1,27 +0,0 @@
-# frozen_string_literal: true
-
-require "digest"
-require "nokogiri"
-require_relative "reverse_adoc/version"
-require_relative "reverse_adoc/errors"
-require_relative "reverse_adoc/cleaner"
-require_relative "reverse_adoc/config"
-require_relative "reverse_adoc/converters"
-require_relative "reverse_adoc/converters/base"
-require_relative "reverse_adoc/html_converter"
-
-module ReverseAdoc
-  def self.convert(input, options = {})
-    ReverseAdoc::HtmlConverter.convert(input, options)
-  end
-
-  def self.config
-    @config ||= Config.new
-    yield @config if block_given?
-    @config
-  end
-
-  def self.cleaner
-    @cleaner ||= Cleaner.new
-  end
-end

data/spec/assets/anchors.html

@@ -1,30 +0,0 @@
-<html>
-  <body>
-    some text...
-    <a href="http://foobar.com">Foobar</a>
-    <a href="http://foobar.com" title="f***** up beyond all recognition">Fubar</a>
-    <a href="http://foobar.com" title="f***** up beyond all redemption"/>
-    <a href="http://strong.foobar.com"><strong>Strong foobar</strong></a>
-    There should be space before but not after the anchor (<a href="http://foobar.com">stripped</a>).
-
-    do not ignore <a href="foo.html">   </a> anchor tags with no link text
-    link <a href="#content">internal jumplinks</a> with anchors
-    link <a href="#content2"/>internal jumplinks without anchors
-    treat <a id="content">anchor tags with no href</a> as bookmarks
-
-    some text...
-
-    <a href="#a__b spaced">Double __ anchor with space</a>
-
-    Double __ anchor with space <a id="a__b spaced"/>
-
-    <a id="_Toc12345"/>
-
-    <img src="http://foobar.com/logo.png">
-    <img alt="foobar image" src="http://foobar.com/foobar.png">
-    <img alt="foobar image 2" title="this is the foobar image 2" src="http://foobar.com/foobar2.png">
-    some text...
-
-    <h1><a id="a__Foreword"></a>Text</h1>
-  </body>
-</html>

data/spec/assets/basic.html

@@ -1,61 +0,0 @@
-<html>
-  <body>
-    plain text
-    <h1>h1</h1>
-    <h1 id="A">h1 with anchor</h1>
-    <h2>h2</h2>
-    <h3>h3</h3>
-    <h4>h4</h4>
-    <h5>h5</h5>
-    <h6>h6</h6>
-
-    <em>em tag content</em>
-    before <em></em> and after empty em tags
-    before <em> </em> and after em tags containing whitespace
-    before <em> <em> <br /> </em> </em> and after em tags containing whitespace
-    <em><em>double em tags</em></em>
-    <p><em><em>double em tags in p tag</em></em></p>
-    a<em> em with leading and trailing </em>whitespace
-    a<em>     
-        em with extra leading and trailing   
-        </em>whitespace
-
-    <strong>strong tag content</strong>
-    before <strong></strong> and after empty strong tags
-    before <strong> </strong> and after strong tags containing whitespace
-    before <strong> <strong> <br /> </strong> </strong> and after strong tags containing whitespace
-    <strong><strong>double strong tags</strong></strong>
-    <p><strong><strong>double strong tags in p tag</strong></strong></p>
-    before
-    <strong>
-        <strong>
-          double strong tags containing whitespace
-        </strong>
-      </strong> after
-    a<strong> strong with leading and trailing </strong>whitespace
-    a<strong>     
-        strong with extra leading and trailing   
-        </strong>whitespace
-
-    constrai<strong>ned</strong> strong
-    constrai<i>ned</i> italic
-    
-    <b>b tag content</b>
-    <i>i tag content</i>
-
-    H<sub>2</sub>O
-    A<sup>2</sup>B
-
-    br tags become double space followed by newline<br/>
-
-    before hr
-    <hr/>
-    after hr
-
-    <div>section 1</div>
-    <div>section 2</div>
-
-    ignore <abbr>abbr</abbr>
-
-  </body>
-</html>

data/spec/assets/code.html

@@ -1,22 +0,0 @@
-<html>
-  <body>
-  <pre>pre block</pre>
-  <code>code block</code>
-  <pre><code>pre code block</code></pre>
-
-	<p>Paragraph with inline <code>code</code> block</p>
-
-	<pre><code>var this;
-this.is("A multi line code block")
-console.log("Yup, it is")
-	</code></pre>
-
-Code with indentation:
-<pre><code>tell application "Foo"
-    beep
-end tell
-</code></pre>
-
-  </body>
-</html>
-

data/spec/assets/escapables.html

@@ -1,17 +0,0 @@
-<html>
-  <body>
-    some text...
-
-    **two asterisks**
-    ***three asterisks***
-    ***three *and*the* asterisc*word and asterisc**multword asterisks***
-    __two underscores__
-    ___three underscores___
-    ___three__underscores___ and another_undersocre
-
-    some text...
-
-    <pre><code>var theoretical_max_infin = 1.0;</code></pre>
-
-  </body>
-</html>

data/spec/assets/external_images.html

@@ -1,35 +0,0 @@
-<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1 plus MathML 2.0//EN" "http://www.w3.org/Math/DTD/mathml2/xhtml-math11-f.dtd">
-<?xml version="1.0" encoding="UTF-8"??><html xmlns="http://www.w3.org/1999/xhtml" style="margin: 0;">
-<!--This file was converted to xhtml by LibreOffice - see http://cgit.freedesktop.org/libreoffice/core/tree/filter/source/xslt for the code.--><head profile="http://dublincore.org/documents/dcmi-terms/">
-<meta http-equiv="Content-Type" content="application/xhtml+xml; charset=utf-8">
-<meta name="DCTERMS.title" content="" xml:lang="en-US">
-<meta name="DCTERMS.language" content="en-US" scheme="DCTERMS.RFC4646">
-<meta name="DCTERMS.source" content="http://xml.openoffice.org/odf2xhtml">
-<meta name="DCTERMS.creator" content="Nick Nicholas">
-<meta name="DCTERMS.issued" content="2019-11-21T08:48:00" scheme="DCTERMS.W3CDTF">
-<meta name="DCTERMS.contributor" content="Nick Nicholas">
-<meta name="DCTERMS.modified" content="2019-11-21T10:01:00" scheme="DCTERMS.W3CDTF">
-<meta name="DCTERMS.provenance" content="" xml:lang="en-US">
-<meta name="DCTERMS.subject" content="," xml:lang="en-US">
-<link rel="schema.DC" href="http://purl.org/dc/elements/1.1/" hreflang="en">
-<link rel="schema.DCTERMS" href="http://purl.org/dc/terms/" hreflang="en">
-<link rel="schema.DCTYPE" href="http://purl.org/dc/dcmitype/" hreflang="en">
-<link rel="schema.DCAM" href="http://purl.org/dc/dcam/" hreflang="en">
-</head>
-<body dir="ltr" style="max-width: 21.001cm; margin: 2.54cm 3.175cm;">
-<h1 class="P3" style="clear: both; color: #2f5496; font-size: 16pt; font-family: Calibri Light; writing-mode: lr-tb; margin: 0.423cm 0 0cm;" align="left ! important">
-<a id="a__Hello" style="margin: 0;"><span style="margin: 0;"></span></a><span class="T1" style="margin: 0;">Hello</span>
-</h1>
-<p class="P1" style="font-size: 12pt; font-family: Calibri; writing-mode: lr-tb; margin: 0;" align="left ! important"> </p>
-<p class="Standard" style="font-size: 12pt; font-family: Calibri; writing-mode: lr-tb; margin: 0;" align="left ! important"><span class="T1" style="margin: 0;">H</span><h1 class="T3" style="vertical-align: super; font-size: 58%; margin: 0;">2</h1><span class="T1" style="margin: 0;">0</span></p>
-<p class="P1" style="font-size: 12pt; font-family: Calibri; writing-mode: lr-tb; margin: 0;" align="left ! important"> </p>
-<!--Next 'div' was a 'text:p'.--><div class="Standard" style="font-size: 12pt; font-family: Calibri; writing-mode: lr-tb; margin: 0;" align="left ! important">
-<!--Next ' 			span' is a draw:frame. 		--><span style="margin: 0;"><math xmlns="http://www.w3.org/1998/Math/MathML" display="block" style="margin: 0;"><mrow style="margin: 0;"><mrow style="margin: 0;"><mrow style="margin: 0;"></mrow><mrow style="margin: 0;"><mi style="margin: 0;">i</mi><mo stretchy="false" style="margin: 0;">=</mo><mn style="margin: 0;">1</mn></mrow></mrow><mrow style="margin: 0;"><mrow style="margin: 0;"></mrow><mi style="margin: 0;">n</mi><msubsup style="margin: 0;"><mi style="margin: 0;">β</mi><mn style="margin: 0;">2</mn><mi style="margin: 0;">i</mi></msubsup></mrow></mrow></math></span>
-</div>
-<div style="clear: both; line-height: 0; width: 0; height: 0; margin: 0; padding: 0;"> </div>
-<!--Next 'div' was a 'text:p'.--><div class="Standard" style="font-size: 12pt; font-family: Calibri; writing-mode: lr-tb; margin: 0;" align="left ! important">
-<a id="_GoBack" style="margin: 0;"></a><!--Next ' 			span' is a draw:frame. 		--><span style="height: 1.764cm; margin: 0cm; padding: 0; border: none; width: 1.764cm; font-size: 12pt; font-family: Calibri; text-align: center; vertical-align: top; background-color: transparent;" class="fr1" id="Εικόνα_2"><img style="height: 1.764cm; width: 1.764cm; margin: 0;" alt="" src="data:image/gif;base64,R0lGODlhZABkAIAAAAAAAP///ywAAAAAZABkAEAC+4yPqcvtD6OctNqLs968XwCGokd24kidYckuKri9cFvKQCzTdG7x+u+y+YDECqqIzByTzKbzCY1Kp9Sq9eoQKrTY1Ev57TJsuLA4Meylz8HTxMz2LOPFOR1ov+v3/L7/DxgoOEj4IcRVeLA2Bue3CEHGF2nYyDaJcWn5+LZ5lamm0tf5MHpWilYJ6INymmiQ52q0EqsBSythews5qwt60wscLDxMXGx8jJysvMzc7PwM7XmYWzh9SPuJmuq43Raq2ooQbpUdMT5Vbn4ela7+fdfu/h4Xv9stfS/vBr+uPa+Zz94/gPvADCRY0FdCfgdxBezSr1+Vh+Iaclu4KoSiKGsSL9YTRu1YSGMji5UkdnJYSpC8nq0M9hJYzF4zo9m8iTOnzigFAAAh/tNUaGlzIGZpbGUgd2FzIGNyZWF0ZWQgYnkNDUdyYXBoaWMgV29ya3Nob3CZIFByb2Zlc3Npb25hbCAyLjBhDQ1mcm9tIEFsY2hlbXkgTWluZHdvcmtzIEluYy4NaHR0cDovL3d3dy5taW5kd29ya3Nob3AuY29tDQ1UaGlzIGltYWdlIG1heSBoYXZlIGJlZW4gY3JlYXRlZCBieQ1hIHBhcnR5IG90aGVyIHRoYW4gQWxjaGVteSBNaW5kd29ya3MgSW5jLg0NVXNlIG5vIGhvb2tzADs="></span>
-</div>
-<div style="clear: both; line-height: 0; width: 0; height: 0; margin: 0; padding: 0;"> </div>
-</body>
-</html>

data/spec/assets/from_the_wild.html

@@ -1,23 +0,0 @@
-<p>
-  <strong>
-    <strong>
-      .<br />
-    </strong>
-    *** intentcast
-  </strong>
-  : logo design
-  <strong>
-    <strong>
-      <br />
-    </strong>
-  </strong>
-  <strong>
-    <strong>
-      .
-    </strong>
-  </strong>
-</p>
-
-<a href="example.com/foo_bar">
-  <img src="example.com/foo_bar.png"> I\_AM\_HELPFUL
-</a>