Allow generating sitemap_index files

lib/jekyll/jekyll-sitemap.rb

@@ -10,6 +10,20 @@ class JekyllSitemap < Jekyll::Generator
     # Main plugin action, called by Jekyll-core
     def generate(site)
       @site = site
+
+      @config = site.config["jekyll_sitemap"]
+      @config = {} unless @config.is_a?(Hash) && @config["index"].is_a?(Hash)
+
+      @index_filename = @config.dig("index", "filename") || "sitemap_index.xml"
+      @index_entries =  @config.dig("index", "linked_sitemaps")
+
+      if @index_entries.is_a?(Array)
+        @priority_sitemap = @index_filename
+        @site.pages << sitemap_index unless file_exists?("sitemap_index.xml")
+      else
+        @priority_sitemap = "sitemap.xml"
+      end
+
       @site.pages << sitemap unless file_exists?("sitemap.xml")
       @site.pages << robots unless file_exists?("robots.txt")
     end
@@ -45,6 +59,16 @@ def destination_path(file = "sitemap.xml")
       @site.in_dest_dir(file)
     end
 
+    def sitemap_index
+      index = PageWithoutAFile.new(@site, __dir__, "", "sitemap_index.xml")
+      index.content = File.read(source_path("sitemap_index.xml")).gsub(MINIFY_REGEX, "")
+      index.data["layout"] = nil
+      index.data["permalink"] = "/#{@index_filename}"
+      index.data["linked_sitemaps"] = @index_entries
+      index.data["xsl"] = file_exists?("sitemap_index.xsl")
+      index
+    end
+
     def sitemap
       site_map = PageWithoutAFile.new(@site, __dir__, "", "sitemap.xml")
       site_map.content = File.read(source_path).gsub(MINIFY_REGEX, "")
@@ -58,6 +82,7 @@ def robots
       robots = PageWithoutAFile.new(@site, __dir__, "", "robots.txt")
       robots.content = File.read(source_path("robots.txt"))
       robots.data["layout"] = nil
+      robots.data["priority_sitemap"] = @priority_sitemap
       robots
     end
 

lib/robots.txt

@@ -1 +1 @@
-Sitemap: {{ "sitemap.xml" | absolute_url }}
+Sitemap: {{ page.priority_sitemap | absolute_url }}

lib/sitemap_index.xml

@@ -0,0 +1,13 @@
+<?xml version="1.0" encoding="UTF-8"?>
+{% if page.xsl %}<?xml-stylesheet type="text/xsl" href="{{ page.xsl | absolute_url }}"?>{% endif %}
+<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+  {%- comment %}The first entry is always the sitemap generated for current site.{% endcomment %}
+  <sitemap>
+    <loc>{{ 'sitemap.xml' | absolute_url }}</loc>
+  </sitemap>
+  {% for entry in page.linked_sitemaps %}
+  <sitemap>
+    <loc>{{ entry | absolute_url }}</loc>
+  </sitemap>
+  {% endfor %}
+</sitemapindex>

spec/jekyll-sitemap_spec.rb

@@ -257,4 +257,122 @@
       end
     end
   end
+
+  describe "Sitemap Index" do
+    let(:custom_config) { {} }
+    let(:config) do
+      Jekyll.configuration(
+        Jekyll::Utils.deep_merge_hashes(
+          overrides, custom_config
+        )
+      )
+    end
+    let(:index_filename) { "sitemap_index.xml" }
+    let(:index_entries) do
+      [
+        "repo1/sitemap.xml",
+        "repo2/sitemap.xml",
+        "repo3/custom-sitemap.xml",
+      ].map { |e| "https://username.github.io/#{e}" }
+    end
+    let(:index_contents) { File.read(dest_dir(index_filename)) }
+    let(:robots_contents) { File.read(dest_dir("robots.txt")) }
+
+    context "with default configuration" do
+      it "does not generate a sitemap_index.xml file" do
+        expect(File.exist?(dest_dir("sitemap_index.xml"))).to_not be_truthy
+      end
+
+      it "generates a sitemap.xml file" do
+        expect(File.exist?(dest_dir("sitemap.xml"))).to be_truthy
+      end
+
+      it "generates a robots.txt file" do
+        expect(File.exist?(dest_dir("robots.txt"))).to be_truthy
+        expect(robots_contents).to match("Sitemap: http://example.org/sitemap.xml")
+      end
+    end
+
+    context "with improper configuration" do
+      let(:custom_config) do
+        {
+          "jekyll_sitemap" => {
+            "index" => "www.example.org/sitemap_index.xml",
+          },
+        }
+      end
+
+      it "does not generate a sitemap_index.xml file" do
+        expect(File.exist?(dest_dir("sitemap_index.xml"))).to_not be_truthy
+      end
+
+      it "generates a sitemap.xml file" do
+        expect(File.exist?(dest_dir("sitemap.xml"))).to be_truthy
+      end
+
+      it "generates a robots.txt file" do
+        expect(File.exist?(dest_dir("robots.txt"))).to be_truthy
+        expect(robots_contents).to match("Sitemap: http://example.org/sitemap.xml")
+      end
+    end
+
+    context "with proper configuration - I" do
+      let(:custom_config) do
+        {
+          "baseurl"        => "bass",
+          "jekyll_sitemap" => {
+            "index" => {
+              "linked_sitemaps" => [],
+            },
+          },
+        }
+      end
+
+      it "generates a sitemap_index.xml file" do
+        expect(File.exist?(dest_dir("sitemap_index.xml"))).to be_truthy
+      end
+
+      it "generates a sitemap.xml file" do
+        expect(File.exist?(dest_dir("sitemap.xml"))).to be_truthy
+      end
+
+      it "generates a robots.txt file" do
+        expect(File.exist?(dest_dir("robots.txt"))).to be_truthy
+        expect(robots_contents).to match("Sitemap: http://example.org/bass/sitemap_index.xml")
+      end
+    end
+
+    context "with proper configuration - II" do
+      let(:index_filename) { "sitemap-index.xml" }
+      let(:custom_config) do
+        {
+          "url"            => "https://username.github.io",
+          "jekyll_sitemap" => {
+            "index" => {
+              "filename"        => index_filename,
+              "linked_sitemaps" => index_entries,
+            },
+          },
+        }
+      end
+
+      it "generates a sitemap-index.xml file" do
+        expect(File.exist?(dest_dir("sitemap_index.xml"))).to_not be_truthy
+        expect(File.exist?(dest_dir("sitemap-index.xml"))).to be_truthy
+
+        expect(index_contents).to match("<loc>https://username.github.io/sitemap.xml</loc>")
+        expect(index_contents).to match("<loc>https://username.github.io/repo1/sitemap.xml</loc>")
+        expect(index_contents).to match("<loc>https://username.github.io/repo3/custom-sitemap.xml</loc>")
+      end
+
+      it "generates a sitemap.xml file" do
+        expect(File.exist?(dest_dir("sitemap.xml"))).to be_truthy
+      end
+
+      it "generates a robots.txt file" do
+        expect(File.exist?(dest_dir("robots.txt"))).to be_truthy
+        expect(robots_contents).to match("Sitemap: https://username.github.io/sitemap-index.xml")
+      end
+    end
+  end
 end