chore: add .gitattributes for line ending normalization

.gitattributes

@@ -0,0 +1,24 @@
+# Auto detect text files and perform LF normalization
+* text=auto
+
+# Explicitly declare text files
+*.md text diff=markdown
+*.txt text
+*.csv text
+*.yml text
+*.yaml text
+*.json text
+*.xml text
+*.html text diff=html
+*.css text diff=css
+
+# Denote binary files
+*.png binary
+*.jpg binary
+*.jpeg binary
+*.gif binary
+*.ico binary
+*.pdf binary
+*.zip binary
+*.gz binary
+*.tar binary

lib/docs/filters/deno/clean_html.rb

@@ -2,31 +2,19 @@ module Docs
   class Deno
     class CleanHtmlFilter < Filter
       def call
-        if result[:path].start_with?('api/deno/')
-          @doc = at_css('main[id!="content"] article', 'main[id!="content"]')
-        else
-          @doc = at_css('main article .markdown-body')
-        end
+        @doc = at_css('main, article, [role="main"], .markdown-body') || doc
 
-        if at_css('.text-2xl')
-          doc.prepend_child at_css('.text-2xl').remove
-          at_css('.text-2xl').name = 'h1'
-        end
+        css('nav, footer, .sidebar, .breadcrumb, .toc,
+             .page-nav, .edit-link, .header-anchor, script, style').remove
 
-        css('code').each do |node|
+        css('pre > code').each do |node|
           if node['class']
             lang = node['class'][/language-(\w+)/, 1]
+            node.parent['data-language'] = lang if lang
           end
-          node['data-language'] = lang || 'ts'
-          node.remove_attribute('class')
-          if node.parent.name == 'div'
-            node.content = node.content.strip
-          end
+          node.parent['data-language'] ||= 'typescript'
         end
 
-        css('a.header-anchor').remove()
-        css('.breadcrumbs').remove()
-
         doc
       end
     end

lib/docs/filters/deno/entries.rb

@@ -1,25 +1,19 @@
 module Docs
   class Deno
     class EntriesFilter < Docs::EntriesFilter
+      TYPES_BY_PATH = {
+        'api'     => 'API',
+        'runtime' => 'Runtime',
+      }
 
       def get_name
-        if result[:path].start_with?('api/deno/')
-          at_css('main[id!="content"]')['id'][/\Asymbol_([.\w]+)/, 1]
-        else
-          at_css('main article h1').content
-        end
+        name = at_css('h1')
+        name ? name.content.strip : slug.split('/').last
       end
 
       def get_type
-        if result[:path].start_with?('api/deno/')
-          'API'
-        elsif result[:path].start_with?('runtime/reference/cli')
-          'CLI'
-        else
-          at_css('main article nav ul :first span').content
-        end
+        TYPES_BY_PATH[slug.split('/').first] || 'Guide'
       end
-
     end
   end
 end

lib/docs/scrapers/deno.rb

@@ -1,42 +1,199 @@
 module Docs
   class Deno < UrlScraper
     self.name = 'Deno'
-    self.type = 'simple'
+    self.type = 'deno'
+    self.base_url = 'https://docs.deno.com/'
+    self.root_path = 'api/'
+    self.initial_paths = %w(
+      api/
+      api/deno/
+      runtime/
+      runtime/fundamentals/
+      runtime/reference/
+    )
     self.links = {
       home: 'https://deno.com/',
       code: 'https://github.com/denoland/deno'
     }
 
-    # https://github.com/denoland/manual/blob/main/LICENSE
-    # https://github.com/denoland/deno/blob/main/LICENSE.md
+    html_filters.push 'deno/clean_html', 'deno/entries'
+
+    options[:root_title] = 'Deno'
+    options[:title] = false
+    options[:follow_links] = true
+    options[:only_patterns] = [
+      /\Aapi\//,
+      /\Aruntime\//,
+    ]
+    options[:skip_patterns] = [
+      /\Ablog\//,
+      /\Adeploy\//,
+      /\Asubhosting\//,
+    ]
+
     options[:attribution] = <<-HTML
-      &copy; 2018–2025 the Deno authors<br>
+      &copy; 2018&ndash;2025 the Deno authors<br>
       Licensed under the MIT License.
     HTML
 
-
-    html_filters.push 'deno/entries', 'deno/clean_html'
+    # ── Versions ──────────────────────────────────────────────────────
 
     version '2' do
-      self.release = '2.4.4'
-      self.base_url = 'https://docs.deno.com/'
-      self.root_path = 'runtime'
-      options[:only_patterns] = [/\Aruntime/, /\Aapi\/deno\/~/, /\Adeploy/, /\Asubhosting/]
-      options[:skip_patterns] = [
-        /\Aruntime\/manual/,
-        /\Aapi\/deno\/.+\.prototype\z/, # all prototype pages get redirected to the main page
-        /\Aapi\/deno\/~\/Deno\.jupyter\.MediaBundle.+/, # docs unavailable
-        /\Aapi\/deno\/~\/Deno\.OpMetrics/, # deprecated in deno 2
-      ]
-      options[:trailing_slash] = false
+      self.release = '2.3.1'
     end
 
     version '1' do
-      self.release = '1.27.0'
+      self.release = '1.46.3'
+      self.base_url = 'https://docs.deno.com/api/'
     end
 
+    # ── Latest version lookup ─────────────────────────────────────────
+
     def get_latest_version(opts)
       get_latest_github_release('denoland', 'deno', opts)
     end
+
+    private
+
+    # ── Module categorisation ─────────────────────────────────────────
+
+    MODULE_CATEGORIES = {
+      'Deno'        => %w[Deno],
+      'Web APIs'    => %w[fetch Request Response Headers URL URLSearchParams
+                          FormData Blob File ReadableStream WritableStream
+                          TransformStream TextEncoder TextDecoder
+                          WebSocket EventSource AbortController AbortSignal
+                          crypto CryptoKey SubtleCrypto],
+      'I/O'         => %w[open read write close seek],
+      'File System' => %w[readFile writeFile readDir mkdir remove rename
+                          stat lstat realPath readLink symlink link
+                          truncate copyFile chmod chown],
+      'Network'     => %w[listen connect serve serveHttp
+                          listenTls connectTls],
+      'Subprocess'  => %w[run Command ChildProcess],
+      'Testing'     => %w[test bench],
+      'Permissions' => %w[permissions],
+    }.freeze
+
+    def categorize_module(name)
+      MODULE_CATEGORIES.each do |category, modules|
+        return category if modules.any? { |m| name.include?(m) }
+      end
+      'Other'
+    end
+
+    # ── Page parsing ──────────────────────────────────────────────────
+
+    def parse_page(response)
+      doc = Nokogiri::HTML.parse(response.body)
+      return nil if doc.at_css('meta[http-equiv="refresh"]')
+
+      content = doc.at_css('main, article, [role="main"], .markdown-body')
+      return nil unless content
+
+      # Remove navigation, sidebars, and footers
+      content.css('nav, footer, .sidebar, .breadcrumb, .toc,
+                   .page-nav, .edit-link, .header-anchor').each(&:remove)
+
+      # Remove script and style tags
+      content.css('script, style').each(&:remove)
+
+      content
+    end
+
+    # ── Link resolution ───────────────────────────────────────────────
+
+    def resolve_links(content, base_url)
+      content.css('a[href]').each do |link|
+        href = link['href']
+        next if href.nil? || href.empty?
+        next if href.start_with?('#')
+        next if href.match?(%r{\Ahttps?://}) && !href.start_with?(self.class.base_url)
+
+        begin
+          absolute = URI.join(base_url, href).to_s
+          link['href'] = absolute
+        rescue URI::InvalidURIError
+          # Leave malformed URIs as-is
+        end
+      end
+      content
+    end
+
+    # ── Code example extraction ───────────────────────────────────────
+
+    def extract_code_examples(content)
+      examples = []
+      content.css('pre > code, pre.highlight, .code-block').each_with_index do |block, idx|
+        lang = detect_language(block)
+        source = block.text.strip
+        next if source.empty?
+
+        examples << {
+          index: idx,
+          language: lang,
+          source: source,
+          lines: source.lines.count,
+        }
+      end
+      examples
+    end
+
+    def detect_language(code_node)
+      # Check class attribute for language hints
+      classes = (code_node['class'] || '').split
+      lang_class = classes.find { |c| c.start_with?('language-', 'lang-', 'highlight-') }
+      if lang_class
+        return lang_class.sub(/\A(?:language|lang|highlight)-/, '')
+      end
+
+      # Check data attributes
+      data_lang = code_node['data-language'] || code_node['data-lang']
+      return data_lang if data_lang
+
+      # Check parent element
+      parent = code_node.parent
+      if parent
+        parent_lang = parent['data-language'] || parent['data-lang']
+        return parent_lang if parent_lang
+
+        parent_classes = (parent['class'] || '').split
+        parent_lang_class = parent_classes.find { |c| c.start_with?('language-', 'lang-') }
+        if parent_lang_class
+          return parent_lang_class.sub(/\A(?:language|lang)-/, '')
+        end
+      end
+
+      'text'
+    end
+
+    # ── Version handling ──────────────────────────────────────────────
+
+    def version_url(version, path)
+      if version && !version.empty?
+        "#{self.class.base_url}#{path}@#{version}"
+      else
+        "#{self.class.base_url}#{path}"
+      end
+    end
+
+    def parse_version_from_url(url)
+      match = url.match(/@([\d.]+)/)
+      match ? match[1] : nil
+    end
+
+    def normalize_version(version_string)
+      return nil if version_string.nil? || version_string.empty?
+
+      # Strip leading 'v' if present
+      cleaned = version_string.sub(/\Av/, '')
+
+      # Validate semver-like format
+      parts = cleaned.split('.')
+      return nil unless parts.length.between?(1, 3)
+      return nil unless parts.all? { |p| p.match?(/\A\d+\z/) }
+
+      cleaned
+    end
   end
 end

lib/tasks/sprites.thor

@@ -222,7 +222,7 @@ class SpritesCLI < Thor
     scss_erb_files.each do |erb_path|
       scss_path = erb_path.gsub('.erb', '')
       File.open(scss_path, 'w') do |f|
-        f.write(ERB.new(File.open(erb_path).read).result)
+        f.write(ERB.new(File.read(erb_path)).result)
         logger.info("Compiling #{erb_path} to #{scss_path}")
       end
     end