Module: Abbreviato

Defined in:
lib/abbreviato/version.rb,
lib/abbreviato/abbreviato.rb

Constant Summary collapse

VERSION = 
'3.1.1'
DEFAULT_OPTIONS = 
{
  max_length: 30,
  tail: '…',
  fragment: true
}.freeze

Class Method Summary collapse

Class Method Details

.truncate(source = '', user_options = {}) ⇒ [String] the truncated string, [boolean] whether the string was truncated

Truncates the source XML string and returns the truncated XML and a boolean flag indicating whether any truncation took place. It will keep a valid XML structure and insert a tail text indicating the position where content was removed (…).

Parameters:

  • source (String) (defaults to: '') —

    the XML source to truncate

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

    truncation options

Options Hash (user_options):

  • :max_length (Integer) —

    Maximum length

  • :truncate_incomplete_row (Boolean) —

    Indicates whether or not to truncate the last row in a table if truncation due to max_length occurs in the middle of a row.

  • :tail (String) —

    Text to append when the truncation happens

  • :fragment (Boolean) —

    Indicates whether the document to be truncated is an HTML fragment or an entire document (with ‘HTML`, `HEAD` & `BODY` tags). Setting to true prevents automatic addition of these tags if they are missing. Defaults to `true`.

Returns:

  • ([String] the truncated string, [boolean] whether the string was truncated) 


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
 
# File 'lib/abbreviato/abbreviato.rb', line 25

def self.truncate(source = '', user_options = {})
  return [nil, false] if source.nil?

  truncated_sax_document = TruncatedSaxDocument.new(DEFAULT_OPTIONS.merge(user_options))
  parser = Nokogiri::HTML::SAX::Parser.new(truncated_sax_document)
  parser.parse(source) { |context| context.replace_entities = false }

  if truncated_sax_document.truncated && user_options[:truncate_incomplete_row]
    parsed_results = [truncated_sax_document.truncated_string.strip, truncated_sax_document.truncated]

    html_fragment = Nokogiri::HTML.fragment(truncated_sax_document.truncated_string.strip)
    return parsed_results if html_fragment.nil?

    last_table_in_doc = html_fragment.xpath('.//table').last
    return parsed_results unless last_table_in_doc

    first_row = last_table_in_doc.xpath('.//tr').first
    return parsed_results unless first_row

    cols_in_first_row = first_row.xpath('.//td').length
    return parsed_results unless cols_in_first_row.positive?

    last_table_in_doc.xpath('.//tr').each do |row|
      row.remove if row.xpath('.//td').length != cols_in_first_row
    end

    return [html_fragment.to_html, truncated_sax_document.truncated]
  end

  [truncated_sax_document.truncated_string.strip, truncated_sax_document.truncated]
end