jekyll-client-search 0.1.0 → 0.3.0

data/lib/jekyll/client_search/document_builder.rb

@@ -5,11 +5,21 @@ module Jekyll
     # Normalizes Jekyll documents and pages into the flat hash shape that the
     # search index JSON emits.
     class DocumentBuilder
-      def from_document(document)
+      def from_document(document, source: nil, passthrough_fields: [])
         url = document.url.to_s
         return if url.empty?
 
         data = document.data
+        core_fields(document, data, url)
+          .compact
+          .merge("source" => source)
+          .merge(passthrough(data, passthrough_fields))
+          .compact
+      end
+
+      private
+
+      def core_fields(document, data, url)
         {
           "id" => url,
           "title" => clean(data["title"] || "Untitled"),
@@ -20,10 +30,18 @@ module Jekyll
           "tags" => normalize_list(data["tags"]),
           "date" => normalized_date(document),
           "date_timestamp" => normalized_timestamp(document)
-        }.compact
+        }
       end
 
-      private
+      def passthrough(data, fields)
+        fields.each_with_object({}) do |(source, target), result|
+          value = data[source]
+          next if value.nil?
+          next if value.is_a?(String) && value.empty?
+
+          result[target] = value
+        end
+      end
 
       def excerpt_for(document)
         document.excerpt if document.respond_to?(:excerpt)
@@ -49,8 +67,8 @@ module Jekyll
 
       def clean(value)
         cleaned = value.to_s
-                       .gsub(%r{<script\b[^>]*>.*?</script>}mi, " ")
-                       .gsub(%r{<style\b[^>]*>.*?</style>}mi, " ")
+                       .gsub(%r{<script\b[^>]*>.*?</script\s*>}mi, " ")
+                       .gsub(%r{<style\b[^>]*>.*?</style\s*>}mi, " ")
                        .gsub(/\{%.*?%\}/m, " ")
                        .gsub(/\{\{.*?\}\}/m, " ")
                        .gsub(/!\[[^\]]*\]\([^)]*\)/, " ")

data/lib/jekyll/client_search/dropdown_configuration.rb

@@ -0,0 +1,92 @@
+# frozen_string_literal: true
+
+require "uri"
+
+module Jekyll
+  module ClientSearch
+    # Validates optional compact dropdown live-search settings.
+    class DropdownConfiguration
+      DEFAULTS = {
+        "enabled" => true,
+        "max_items" => 5,
+        "min_chars" => 2,
+        "debounce_ms" => 150,
+        "redirect_url" => "/search/"
+      }.freeze
+
+      def initialize(config)
+        unless config.is_a?(Hash)
+          raise Jekyll::Errors::FatalException,
+                "client_search dropdown configuration must be a mapping"
+        end
+
+        @values = DEFAULTS.merge(config)
+        validate!
+      end
+
+      def enabled?
+        @values.fetch("enabled") != false
+      end
+
+      def max_items
+        @values.fetch("max_items")
+      end
+
+      def min_chars
+        @values.fetch("min_chars")
+      end
+
+      def debounce_ms
+        @values.fetch("debounce_ms")
+      end
+
+      def redirect_url
+        @values.fetch("redirect_url").to_s
+      end
+
+      def to_h
+        {
+          "enabled" => enabled?,
+          "maxItems" => max_items,
+          "minChars" => min_chars,
+          "debounceMs" => debounce_ms,
+          "redirectUrl" => redirect_url
+        }
+      end
+
+      private
+
+      def validate!
+        validate_boolean!("enabled")
+        validate_integer!("max_items", minimum: 1)
+        validate_integer!("min_chars", minimum: 0)
+        validate_integer!("debounce_ms", minimum: 0)
+        if redirect_url.empty?
+          raise Jekyll::Errors::FatalException,
+                "client_search dropdown redirect_url must not be empty"
+        end
+
+        uri = URI.parse(redirect_url)
+        return if uri.scheme.nil? && uri.host.nil? && !redirect_url.start_with?("//")
+
+        raise Jekyll::Errors::FatalException,
+              "client_search dropdown redirect_url must be a relative path"
+      end
+
+      def validate_boolean!(key)
+        return if [true, false].include?(@values[key])
+
+        raise Jekyll::Errors::FatalException,
+              "client_search dropdown.#{key} must be true or false"
+      end
+
+      def validate_integer!(key, minimum:)
+        value = @values.fetch(key)
+        return if value.is_a?(Integer) && value >= minimum
+
+        raise Jekyll::Errors::FatalException,
+              "client_search dropdown.#{key} must be an integer greater than or equal to #{minimum}"
+      end
+    end
+  end
+end

data/lib/jekyll/client_search/dropdown_tag.rb

@@ -0,0 +1,115 @@
+# frozen_string_literal: true
+
+require "cgi"
+
+module Jekyll
+  module ClientSearch
+    # Liquid tag that renders a compact live-search dropdown suitable for
+    # navbars and headers. Framework-agnostic — emits semantic HTML with
+    # data attributes, no CSS classes from any framework.
+    #
+    #   {% search_dropdown %}
+    #   {% search_dropdown max:10 %}
+    #   {% search_dropdown scripts_only %}
+    #   {% search_dropdown no_scripts %}
+    #
+    # When client_search is disabled the tag renders nothing.
+    class DropdownTag < Liquid::Tag
+      SYNTAX = /\A(max:(\d+))?\s*(scripts_only|no_scripts)?\z/
+
+      def initialize(tag_name, markup, tokens)
+        super
+        @markup = markup.to_s.strip
+        unless (match = @markup.match(SYNTAX))
+          raise Liquid::SyntaxError,
+                "search_dropdown: invalid syntax. Use {% search_dropdown %}, " \
+                "{% search_dropdown max:10 %}, {% search_dropdown scripts_only %}, " \
+                "or {% search_dropdown no_scripts %}"
+        end
+
+        @max_items = match[2].to_i if match[2]
+        raise Liquid::SyntaxError, "search_dropdown: max must be greater than zero" if @max_items && @max_items < 1
+
+        @mode = match[3] || "full"
+        @input_id = "cs-dropdown-input-#{object_id}"
+        @results_id = "cs-dropdown-results-#{object_id}"
+      end
+
+      def render(context)
+        site = context.registers[:site]
+        return "" unless enabled?(site)
+
+        configuration = Configuration.new(site)
+        return "" unless configuration.dropdown_enabled?
+
+        render_mode(configuration, site)
+      end
+
+      private
+
+      def enabled?(site)
+        return false if site.nil?
+
+        config_hash = site.config.fetch("client_search", {})
+        config_hash != false && config_hash["enabled"] != false
+      end
+
+      def render_mode(configuration, site)
+        baseurl = site.config["baseurl"].to_s.gsub(%r{\A/+|/+$}, "")
+        prefix = baseurl.empty? ? "" : "/#{baseurl}"
+        max = resolve_max(configuration)
+
+        form_html = build_form(max)
+        return form_html if @mode == "no_scripts"
+
+        scripts = build_scripts(configuration, prefix)
+        return scripts if @mode == "scripts_only"
+
+        "#{form_html}\n#{scripts}"
+      end
+
+      def resolve_max(configuration)
+        @max_items || configuration.dropdown_max_items
+      end
+
+      def build_form(max)
+        <<~HTML
+          <div class="client-search-dropdown" data-client-search-dropdown>
+            <form role="search" autocomplete="off" data-cs-dropdown-form>
+              <label class="sr-only" for="#{@input_id}">Search</label>
+              <input id="#{@input_id}" type="search" name="q"
+                     placeholder="Search…" autocomplete="off"
+                     aria-expanded="false" aria-controls="#{@results_id}"
+                     data-cs-dropdown-input>
+            </form>
+            <ul id="#{@results_id}" role="listbox" aria-hidden="true"
+                data-cs-dropdown-results data-max-items="#{max}"></ul>
+          </div>
+        HTML
+      end
+
+      def build_scripts(configuration, prefix)
+        scripts = []
+        engine = configuration.engine_url
+        if engine
+          attrs = ["src=\"#{CGI.escapeHTML(engine)}\""]
+          if configuration.engine_crossorigin
+            crossorigin = CGI.escapeHTML(configuration.engine_crossorigin)
+            attrs << "crossorigin=\"#{crossorigin}\""
+          end
+          if configuration.engine_sri
+            integrity = CGI.escapeHTML(configuration.engine_sri)
+            attrs << "integrity=\"#{integrity}\""
+          end
+          scripts << "<script #{attrs.join(' ')}></script>"
+        end
+        scripts << "<script src=\"#{prefix}/assets/search-runtime-config.js\"></script>"
+        scripts << "<script src=\"#{prefix}/assets/client-search-dropdown.js\"></script>"
+        scripts << "<script src=\"#{prefix}/assets/adapters/#{configuration.engine}.js\"></script>"
+        scripts.map { |script| "  #{script}" }.join("\n")
+      end
+    end
+  end
+end
+
+Liquid::Template.register_tag("search_dropdown", Jekyll::ClientSearch::DropdownTag)

data/lib/jekyll/client_search/generator.rb

@@ -90,19 +90,23 @@ module Jekyll
       end
 
       def collection_documents(site, configuration, builder)
+        fields = configuration.passthrough_fields
         configuration.collections.flat_map do |label|
           collection = label == "posts" ? site.posts : site.collections[label]
           next [] unless collection
 
-          collection.docs.filter_map { |document| builder.from_document(document) }
+          collection.docs.filter_map do |document|
+            builder.from_document(document, source: label, passthrough_fields: fields)
+          end
         end
       end
 
       def page_documents(site, builder, configuration)
+        fields = configuration.passthrough_fields
         site.pages
             .reject { |page| page.url == "/#{configuration.output}" }
             .select { |page| page.data["title"] }
-            .filter_map { |page| builder.from_document(page) }
+            .filter_map { |page| builder.from_document(page, source: "pages", passthrough_fields: fields) }
       end
 
       def add_runtime_asset(site, configuration)

data/lib/jekyll/client_search/related_configuration.rb

@@ -12,7 +12,7 @@ module Jekyll
         "include_parent_domains" => true,
         "semantic" => true,
         "minimum_similarity" => 0.55,
-        "max_items" => nil
+        "max_items" => 5
       }.freeze
 
       def initialize(config)

data/lib/jekyll/client_search/related_tag.rb

@@ -7,35 +7,45 @@ module Jekyll
     #
     #   {% related_articles %}
     #   {% related_articles sort:date %}
+    #   {% related_articles max:3 %}
+    #   {% related_articles sort:date max:3 %}
     #   {% related_articles no_scripts %}
     #
     # When +related.enabled+ is false the tag renders nothing, so it is safe
     # to leave in a layout even when the feature is off.
     class RelatedTag < Liquid::Tag
-      SYNTAX = /\A\s*(sort:(\w+))?\s*(no_scripts)?\s*\z/
+      SYNTAX = /\A(sort:(\w+))?\s*(max:(\d+))?\s*(no_scripts)?\z/
 
       def initialize(tag_name, markup, tokens)
         super
-        @markup = markup.to_s
+        @markup = markup.to_s.strip
         unless (match = @markup.match(SYNTAX))
           raise Liquid::SyntaxError,
                 "related_articles: invalid syntax. Use {% related_articles %}, " \
-                "{% related_articles sort:date %}, or {% related_articles no_scripts %}"
+                "{% related_articles sort:date %}, {% related_articles max:3 %}, " \
+                "or {% related_articles no_scripts %}"
         end
 
         @sort = match[2] if match[2]
-        @include_scripts = match[3].nil?
+        @max_items = match[4].to_i if match[4]
+        raise Liquid::SyntaxError, "related_articles: max must be greater than zero" if @max_items && @max_items < 1
+
+        @include_scripts = match[5].nil?
       end
 
       def render(context)
         site = context.registers[:site]
-        config = site&.config&.fetch("client_search", {})
+        return "" if site.nil?
+
+        config = site.config.fetch("client_search", {})
+        return "" if config == false
         return "" unless related_enabled?(config)
 
         asset_prefix = asset_prefix(site)
         sort_attr = @sort ? " data-related-sort=\"#{@sort}\"" : ""
+        max_attr = resolve_max_attr(config)
         scripts = build_scripts(asset_prefix)
-        build_html(sort_attr, scripts)
+        build_html(sort_attr, max_attr, scripts)
       end
 
       private
@@ -57,7 +67,19 @@ module Jekyll
           "<script src=\"#{prefix}/assets/client-search-related.js\"></script>"
       end
 
-      def build_html(sort_attr, scripts)
+      def resolve_max_attr(config)
+        related = config.fetch("related", {})
+        max = if @max_items
+                @max_items
+              elsif related.key?("max_items")
+                related["max_items"]
+              else
+                5
+              end
+        max.nil? ? "" : " data-related-max=\"#{max}\""
+      end
+
+      def build_html(sort_attr, max_attr, scripts)
         <<~HTML
           <section class="related-articles-section">
             <label for="related-sort">Sort related articles</label>
@@ -65,7 +87,7 @@ module Jekyll
               <option value="relevance">Most related</option>
               <option value="date">Newest</option>
             </select>
-            <div id="related-articles"#{sort_attr}></div>
+            <div id="related-articles"#{sort_attr}#{max_attr}></div>
           </section>#{scripts}
         HTML
       end

data/lib/jekyll/client_search/runtime_config_page.rb

@@ -12,10 +12,14 @@ module Jekyll
           "liveSearch" => configuration.live_search_config
         }
         defaults["relatedUrl"] = index_url(site, configuration.related_output) if configuration.related_enabled?
+        defaults["iconField"] = configuration.runtime_icon_field if configuration.runtime_icon_field
+        defaults["dropdown"] = configuration.dropdown_config if configuration.dropdown_enabled?
         json = JSON.generate(defaults)
         self.content = "window.clientSearchConfig = (function (generated, existing) {" \
                        "var liveSearch = Object.assign({}, generated.liveSearch, existing.liveSearch || {});" \
-                       "return Object.assign(generated, existing, { liveSearch: liveSearch });" \
+                       "var dropdown = Object.assign({}, generated.dropdown, existing.dropdown || {});" \
+                       "return Object.assign(generated, existing, " \
+                       "{ liveSearch: liveSearch, dropdown: dropdown });" \
                        "}(#{json}, window.clientSearchConfig || {}));\n"
       end
 

data/lib/jekyll/client_search/search_tag.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "cgi"
+
 module Jekyll
   module ClientSearch
     # Liquid tag that renders the search form, status/results containers,
@@ -12,11 +14,11 @@ module Jekyll
     #
     # When client_search is disabled the tag renders nothing.
     class SearchTag < Liquid::Tag
-      SYNTAX = /\A\s*(scripts_only|no_scripts)?\s*\z/
+      SYNTAX = /\A(scripts_only|no_scripts)?\z/
 
       def initialize(tag_name, markup, tokens)
         super
-        @markup = markup.to_s
+        @markup = markup.to_s.strip
         unless (match = @markup.match(SYNTAX))
           raise Liquid::SyntaxError,
                 "search_form: invalid syntax. Use {% search_form %}, " \
@@ -28,7 +30,10 @@ module Jekyll
 
       def render(context)
         site = context.registers[:site]
-        config_hash = site&.config&.fetch("client_search", {})
+        return "" if site.nil?
+
+        config_hash = site.config.fetch("client_search", {})
+        return "" if config_hash == false
         return "" unless config_hash["enabled"] != false
 
         configuration = Configuration.new(site)
@@ -73,9 +78,15 @@ module Jekyll
         url = configuration.engine_url
         return nil unless url
 
-        attrs = ["src=\"#{url}\""]
-        attrs << "crossorigin=\"#{configuration.engine_crossorigin}\"" if configuration.engine_crossorigin
-        attrs << "integrity=\"#{configuration.engine_sri}\"" if configuration.engine_sri
+        attrs = ["src=\"#{CGI.escapeHTML(url)}\""]
+        if configuration.engine_crossorigin
+          crossorigin = CGI.escapeHTML(configuration.engine_crossorigin)
+          attrs << "crossorigin=\"#{crossorigin}\""
+        end
+        if configuration.engine_sri
+          integrity = CGI.escapeHTML(configuration.engine_sri)
+          attrs << "integrity=\"#{integrity}\""
+        end
         "<script #{attrs.join(' ')}></script>"
       end
 
@@ -84,7 +95,13 @@ module Jekyll
 
         scripts = ["<script src=\"#{prefix}/assets/search-embedder-config.js\"></script>"]
         embedder_asset = configuration.query_embedder_asset
+        # simplecov:disable branch
+        # Defensive guard: semantic_with_embedder? already excludes the "none"
+        # type, and both "transformers" and "ollama_api" always return an asset.
+        # This branch is unreachable with current embedder types but protects
+        # against future types that may have no script asset.
         scripts << "<script src=\"#{prefix}/#{embedder_asset}\"></script>" if embedder_asset
+        # simplecov:enable branch
         scripts
       end
 

data/lib/jekyll/client_search/version.rb

@@ -2,6 +2,6 @@
 
 module Jekyll
   module ClientSearch
-    VERSION = "0.1.0"
+    VERSION = "0.3.0"
   end
 end

data/lib/jekyll/client_search.rb

@@ -9,6 +9,7 @@ require_relative "client_search/version"
 require_relative "client_search/embedding_configuration"
 require_relative "client_search/live_search_configuration"
 require_relative "client_search/related_configuration"
+require_relative "client_search/dropdown_configuration"
 require_relative "client_search/query_embedder_configuration"
 require_relative "client_search/configuration_accessors"
 require_relative "client_search/configuration"
@@ -22,4 +23,5 @@ require_relative "client_search/index_cache"
 require_relative "client_search/ollama_embedding_adapter"
 require_relative "client_search/generator"
 require_relative "client_search/related_tag"
+require_relative "client_search/dropdown_tag"
 require_relative "client_search/search_tag"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-client-search
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Svend Gundestrup
@@ -46,6 +46,7 @@ files:
 - assets/adapters/minisearch.js
 - assets/adapters/semantic.js
 - assets/client-search-base.js
+- assets/client-search-dropdown.js
 - assets/client-search-related.js
 - assets/includes/related-articles.html
 - assets/layouts/post-with-related.html
@@ -59,6 +60,8 @@ files:
 - lib/jekyll/client_search/configuration.rb
 - lib/jekyll/client_search/configuration_accessors.rb
 - lib/jekyll/client_search/document_builder.rb
+- lib/jekyll/client_search/dropdown_configuration.rb
+- lib/jekyll/client_search/dropdown_tag.rb
 - lib/jekyll/client_search/embedder_config_page.rb
 - lib/jekyll/client_search/embedding_configuration.rb
 - lib/jekyll/client_search/generator.rb