Class: Jekyll::L10n::UrlTransformer

Inherits:
Object
  • Object
show all
Defined in:
lib/jekyll-l10n/utils/url_transformer.rb

Overview

Transforms relative URLs in HTML to include locale prefixes.

UrlTransformer modifies href attributes in links to prefix them with the target locale in BCP 47 hyphenated form (e.g., /docs/page.html becomes /es-ES/docs/page.html for a canonical "es_ES" locale). It preserves external links, anchors, mailto, tel links, and links already containing locale prefixes. Skips English locale (default language).

Key responsibilities:

  • Identify relative links to transform
  • Add locale prefix in BCP 47 form to href values
  • Preserve external links and special URLs
  • Skip already-localized URLs
  • Handle baseurl paths correctly
  • Remove auto-inserted meta tags

Examples:

html = '<a href="/docs/page.html">Link</a>'
transformed = UrlTransformer.transform(html, 'es_ES', '/baseurl')
# Returns '<a href="/baseurl/es-ES/docs/page.html">Link</a>'

Class Method Summary collapse

Class Method Details

.transform(html, locale, baseurl) ⇒ String

Transform all relative URLs in HTML to include locale prefix.

Parses HTML document, identifies relative links, adds locale prefix in BCP 47 hyphenated form, removes auto-inserted meta tags, and returns transformed HTML.

Parameters:

  • html (String) —

    HTML content with URLs to transform

  • locale (String) —

    Target locale code (e.g., 'es', 'es_ES', 'fr')

  • baseurl (String) —

    Base URL for site (e.g., '/baseurl')

Returns:

  • (String) —

    HTML with locale-prefixed URLs



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
 
# File 'lib/jekyll-l10n/utils/url_transformer.rb', line 42

def transform(html, locale, baseurl)
  # Convert canonical locale to BCP 47 hyphenated form for URL usage.
  # This is idempotent: already-hyphenated input is unchanged.
  locale = LocaleUrlFormatter.to_url_segment(locale)

  return html if should_skip_transform?(locale)

  # Use Nokogiri::HTML to properly parse full HTML documents while preserving
  # DOCTYPE, html tag, and document structure. Auto-inserted meta tags are
  # removed via regex post-processing (same approach as HtmlTranslator).
  # See: spec/regression/nokogiri_meta_tag_spec.rb for regression tests
  doc = Nokogiri::HTML(html)

  transform_document(doc, locale, baseurl)

  result = doc.to_html

  HtmlParser.remove_meta_charset(result)
end

.transform_document(doc, locale, baseurl) ⇒ void

This method returns an undefined value.

Transform URLs in a parsed HTML document.

Modifies href attributes of links in the parsed document in place. Useful when document is already parsed to avoid re-parsing.

Parameters:

  • doc (Nokogiri::HTML::Document) —

    Parsed HTML document

  • locale (String) —

    Target locale code (canonical or BCP 47 form)

  • baseurl (String) —

    Base URL for site



71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/jekyll-l10n/utils/url_transformer.rb', line 71

def transform_document(doc, locale, baseurl)
  # Convert canonical locale to BCP 47 hyphenated form for URL usage.
  # This is idempotent: already-hyphenated input is unchanged.
  locale = LocaleUrlFormatter.to_url_segment(locale)

  return if should_skip_transform?(locale)

  doc.css('a[href]').each do |link|
    href = link['href']
    next unless should_transform_href?(href, locale, baseurl)

    next if language_dropdown_link?(link)

    link['href'] = add_locale_prefix(href, locale, baseurl)
  end
end