Class: Coradoc::Input::Html::HtmlConverter

Inherits:
Object
  • Object
show all
Defined in:
lib/coradoc/html/input/html_converter.rb

Overview

HTML to CoreModel converter

This class handles the conversion of HTML documents to CoreModel. It does NOT handle serialization to any specific output format. For serialization, use Coradoc.serialize(coremodel, to: :format)

Examples:

Basic usage - get CoreModel

coremodel = HtmlConverter.to_core_model(html_string)

Serialize to AsciiDoc

coremodel = HtmlConverter.to_core_model(html_string)
adoc_text = Coradoc.serialize(coremodel, to: :asciidoc)

Class Method Summary collapse

Class Method Details

.cleanup_result(result, options = {}) ⇒ String

Clean up the serialized result

Parameters:

  • result (String) —

    Serialized result

  • options (Hash) (defaults to: {}) —

    Cleanup options

Returns:

  • (String) —

    Cleaned result



125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
 
# File 'lib/coradoc/html/input/html_converter.rb', line 125

def self.cleanup_result(result, options = {})
  Input::Html.config.with(options) do
    plugin_instances = prepare_plugin_instances(options)

    result = track_time 'Cleaning up the result' do
      Input::Html.cleaner.tidy(result)
    end

    plugin_instances.each do |plugin|
      next unless plugin.public_methods.include?(:postprocess_output_string)

      plugin.output_string = result
      track_time "Postprocessing output string with #{plugin.name} plugin" do
        plugin.postprocess_output_string
      end
      result = plugin.output_string
    end

    result
  end
end

.convert(input, options = {}) ⇒ String

Deprecated.

Use #to_core_model + Coradoc.serialize instead

Legacy method for backward compatibility Converts HTML to CoreModel, then serializes to target format

Parameters:

  • input (String) —

    HTML input

  • options (Hash) (defaults to: {}) —

    Conversion options

  • options (Symbol) (defaults to: {}) —

    :output_format Target format (default: :asciidoc)

Returns:

  • (String) —

    Serialized document in target format



93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
 
# File 'lib/coradoc/html/input/html_converter.rb', line 93

def self.convert(input, options = {})
  output_format = options.delete(:output_format) || :asciidoc

  coremodel = to_core_model(input, options)

  if coremodel.is_a?(Hash)
    coremodel.to_h do |file, tree|
      track_time "Serializing file #{file || 'main'}" do
        [file, serialize_core_model(tree, output_format, options)]
      end
    end
  else
    serialize_core_model(coremodel, output_format, options)
  end
end

.prepare_plugin_instances(options) ⇒ Object 



147
148
149
 
# File 'lib/coradoc/html/input/html_converter.rb', line 147

def self.prepare_plugin_instances(options)
  options[:plugin_instances] || Html.config.plugins.map(&:new)
end

.serialize_core_model(coremodel, format, options = {}) ⇒ String

Serialize CoreModel to target format using the appropriate gem

Parameters:

  • coremodel (Coradoc::CoreModel::Base) —

    CoreModel document

  • format (Symbol) —

    Target format

  • options (Hash) (defaults to: {}) —

    Serialization options

Returns:

  • (String) —

    Serialized document



115
116
117
118
 
# File 'lib/coradoc/html/input/html_converter.rb', line 115

def self.serialize_core_model(coremodel, format, options = {})
  result = Coradoc.serialize(coremodel, to: format)
  cleanup_result(result, options)
end

.to_coradoc(input, options = {}) ⇒ Object

Deprecated.

Use #to_core_model instead

Legacy method - returns CoreModel



81
82
83
 
# File 'lib/coradoc/html/input/html_converter.rb', line 81

def self.to_coradoc(input, options = {})
  to_core_model(input, options)
end

.to_core_model(input, options = {}) ⇒ Coradoc::CoreModel::Base

Convert HTML to CoreModel

Parameters:

  • input (String, Nokogiri::XML::Document, Nokogiri::XML::Node) —

    HTML input

  • options (Hash) (defaults to: {}) —

    Conversion options

Returns:

  • (Coradoc::CoreModel::Base) —

    CoreModel document



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
 
# File 'lib/coradoc/html/input/html_converter.rb', line 25

def self.to_core_model(input, options = {})
  Input::Html.config.with(options) do
    plugin_instances = prepare_plugin_instances(options)

    root = track_time 'Loading input HTML document' do
      case input
      when String
        Nokogiri::HTML(input).root
      when Nokogiri::XML::Document
        input.root
      when Nokogiri::XML::Node
        input
      end
    end

    return nil unless root

    plugin_instances.each do |plugin|
      plugin.html_tree = root
      if plugin.public_methods.include?(:preprocess_html_tree)
        track_time "Preprocessing document with #{plugin.name} plugin" do
          plugin.preprocess_html_tree
        end
      end
      root = plugin.html_tree
    end

    coremodel = track_time 'Converting input document tree to CoreModel' do
      Converters.process_coradoc(
        root,
        plugin_instances: plugin_instances
      )
    end

    coremodel = track_time 'Post-process CoreModel tree' do
      Postprocessor.process(coremodel)
    end

    plugin_instances.each do |plugin|
      next unless plugin.public_methods.include?(:postprocess_coremodel_tree)

      plugin.coremodel_tree = coremodel
      track_time "Postprocessing CoreModel tree with #{plugin.name} plugin" do
        plugin.postprocess_coremodel_tree
      end
      coremodel = plugin.coremodel_tree
    end

    options[:plugin_instances] = plugin_instances unless options.frozen?

    coremodel
  end
end

.track_time(task) ⇒ Object 



152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
 
# File 'lib/coradoc/html/input/html_converter.rb', line 152

def self.track_time(task)
  if Input::Html.config.track_time
    warn ('  ' * @track_time_indentation) + "* #{task} is starting..."
    @track_time_indentation += 1
    t0 = Time.now
    ret = yield
    time_elapsed = Time.now - t0
    @track_time_indentation -= 1
    warn ('  ' * @track_time_indentation) +
         "* #{task} took #{time_elapsed.round(3)} seconds"
    ret
  else
    yield
  end
end