Merge pull request #1738 from thatsmydoing/nix

Add nix and nixpkgs

assets/javascripts/news.json

@@ -1,4 +1,8 @@
 [
+  [
+    "2022-04-25",
+    "New documentation: <a href=\"/nix/\">Nix</a>"
+  ],
   [
     "2022-03-31",
     "New documentation: <a href=\"/eigen3/\">Eigen3</a>"

assets/javascripts/templates/pages/about_tmpl.coffee

@@ -587,6 +587,11 @@ credits = [
     '2006-2021 Andreas Rumpf',
     'MIT',
     'https://github.com/nim-lang/Nim#license'
+  ], [
+    'Nix',
+    '2022 NixOS Contributors',
+    'LGPLv2.1',
+    'https://github.com/NixOS/nix#license'
   ], [
     'Node.js',
     'Joyent, Inc. and other Node contributors<br>Node.js is a trademark of Joyent, Inc.',

lib/docs/filters/nix/clean_html.rb

@@ -0,0 +1,121 @@
+module Docs
+  class Nix
+    class CleanHtmlFilter < Filter
+      def call
+        if subpath == 'nixpkgs/stable/index.html'
+          new_root = Nokogiri::XML::Node.new 'div', doc.document
+
+          # lib functions
+          lib_sections = xpath("//*[@id='sec-functions-library']/ancestor::section[1]/section/section")
+          lib_sections.css('.titlepage .title').each do |title|
+            title.name = 'h2'
+            strip_section_number(title.children)
+            title['id'] = title.content.gsub(/[^a-zA-Z0-9]/, '-')
+          end
+          lib_sections.css('.example .title strong').each do |title|
+            title.content = title.content.sub(/^Example ([0-9.]+)/, 'Example: ')
+          end
+          new_root.add_child(lib_sections)
+
+          # fetchers
+          fetcher_sections = xpath("//*[@id='chap-pkgs-fetchers']/ancestor::section[1]/section[position()>1]")
+          fetcher_sections.css('.titlepage .title').each do |title|
+            strip_section_number(title.children)
+            prefix_with(title, 'pkgs') if title.name == 'h2'
+          end
+          new_root.add_child(fetcher_sections)
+
+          # trivial builders
+          trivial_sections = xpath("//*[@id='chap-trivial-builders']/ancestor::section[1]/section")
+          trivial_sections.css('.titlepage .title').each do |title|
+            strip_section_number(title.children)
+            prefix_with(title, 'pkgs') if title.name == 'h2'
+          end
+          new_root.add_child(trivial_sections)
+
+          # special builders
+          special_sections = xpath("//*[@id='chap-special']/ancestor::section[1]/section")
+          special_sections.css('.titlepage .title').each do |title|
+            strip_section_number(title.children)
+            if title.name == 'h2'
+              title.children[0].wrap('<code>')
+              prefix_with(title, 'pkgs')
+            end
+          end
+          new_root.add_child(special_sections)
+
+          # image builders
+          image_sections = xpath("//*[@id='chap-images']/ancestor::section[1]/section")
+          image_sections.css('.titlepage .title').each do |title|
+            strip_section_number(title.children)
+            title.children[0].wrap('<code>')
+          end
+          image_sections.each do |section|
+            prefix = section.at_xpath('*[@class="titlepage"]//*[@class="title"]').content
+            next unless ["pkgs.dockerTools", "pkgs.ociTools"].include?(prefix)
+            section.xpath('section/*[@class="titlepage"]//*[@class="title"]').each do |title|
+              prefix_with(title, prefix)
+              title['data-add-to-index'] = ''
+            end
+          end
+          new_root.add_child(image_sections)
+
+          new_root.css('pre.programlisting').attr('data-language', 'nix')
+
+          new_root
+        elsif subpath == 'nix/stable/expressions/builtins.html'
+          @doc = doc.at_css('main dl')
+
+          # strip out the first entry, `derivation`, the actual documentation
+          # exists in a separate page
+          derivation_dt = doc.children.at_css('dt')
+          if derivation_dt.content.starts_with?('derivation')
+            derivation_dt.remove
+            doc.children.at_css('dd').remove
+          else
+            raise RuntimeError.new('First entry is not derivation, update the scraper')
+          end
+
+          doc.css('dt').each do |title|
+            title.name = 'h2'
+            unwrap(title.at_css('a'))
+            title.children[0].children.before('builtins.')
+          end
+          doc.css('dd').each do |description|
+            description.name = 'div'
+          end
+
+          doc.css('pre > code').each do |code|
+            code.parent['data-language'] = 'nix' if code['class'] == 'language-nix'
+            code.parent['data-language'] = 'xml' if code['class'] == 'language-xml'
+            unwrap(code)
+          end
+
+          doc
+        else
+          doc
+        end
+      end
+
+      def strip_section_number(title_children)
+        while title_children.first.content == ''
+          title_children.shift.remove
+        end
+        first_text = title_children.first
+        return unless first_text.text?
+        first_text.content = first_text.content.sub(/[0-9.]+ ?/, '')
+        first_text.remove if first_text.blank?
+      end
+
+      def prefix_with(title_node, text)
+        title_node.css('code').each do |code|
+          code.content = "#{text}.#{code.content}" unless code.content.starts_with?("#{text}.")
+        end
+      end
+
+      def unwrap(node)
+        node.replace(node.inner_html)
+      end
+    end
+  end
+end

lib/docs/filters/nix/entries.rb

@@ -0,0 +1,22 @@
+module Docs
+  class Nix
+    class EntriesFilter < Docs::EntriesFilter
+      def include_default_entry?
+        false
+      end
+
+      def additional_entries
+        css('h2, h3[data-add-to-index]').flat_map do |node|
+          node.css('code').map do |code|
+            title = code.content
+            index = title.rindex('.')
+            type = title[0...index]
+            name = title.match(/^[^\s]+/)[0]
+
+            [name, node['id'], type]
+          end
+        end
+      end
+    end
+  end
+end

lib/docs/scrapers/nix.rb

@@ -0,0 +1,31 @@
+module Docs
+  class Nix < UrlScraper
+    self.type = 'simple'
+    self.release = '2.8.0'
+    self.base_url = 'https://nixos.org/manual/'
+    self.root_path = 'nix/stable/expressions/builtins.html'
+    self.initial_paths = %w(
+      nix/stable/expressions/builtins.html
+      nixpkgs/stable/index.html)
+    self.links = {
+      home: 'https://nixos.org/',
+      code: 'https://github.com/NixOS/nix'
+    }
+
+    html_filters.push 'nix/clean_html', 'nix/entries'
+
+    options[:skip_links] = true
+
+    options[:attribution] = <<-HTML
+      &copy; 2022 NixOS Contributors<br>
+      Licensed under the LGPL License.
+    HTML
+
+    def get_latest_version(opts)
+      doc = fetch_doc('https://nixos.org/manual/nix/stable/', opts)
+      json = JSON.parse(doc.at_css('body')['data-nix-channels'])
+      channel = json.find { |c| c['channel'] == 'stable' }
+      channel['version']
+    end
+  end
+end

public/icons/docs/nix/SOURCE

@@ -0,0 +1 @@
+https://github.com/NixOS/nixos-homepage/blob/master/logo/nixos-logo-only-hires.png