Split sitemap (#619)

Split sitemap when it is getting too big

.github/ISSUE_TEMPLATE/documentation.md

@@ -10,5 +10,5 @@ What is the issue? Is the documentation unclear? Is it missing information?
 
 ## Proposed solution
 A quick explanation of what you would like to see to solve the issue.
-If you want to add content, please explain what you were looking fod and what was
+If you want to add content, please explain what you were looking for and what was
 your process while looking at the current documentation.

components/rebuild/src/lib.rs

@@ -369,6 +369,7 @@ pub fn after_template_change(site: &mut Site, path: &Path) -> Result<()> {
     match filename {
         "sitemap.xml" => site.render_sitemap(),
         "rss.xml" => site.render_rss_feed(site.library.read().unwrap().pages_values(), None),
+        "split_sitemap_index.xml" => site.render_sitemap(),
         "robots.txt" => site.render_robots(),
         "single.html" | "list.html" => site.render_taxonomies(),
         "page.html" => {

components/site/src/lib.rs

@@ -788,8 +788,6 @@ impl Site {
     pub fn render_sitemap(&self) -> Result<()> {
         ensure_directory_exists(&self.output_path)?;
 
-        let mut context = Context::new();
-
         let mut pages = self
             .library
             .read()
@@ -806,7 +804,6 @@ impl Site {
             })
             .collect::<Vec<_>>();
         pages.sort_by(|a, b| a.permalink.cmp(&b.permalink));
-        context.insert("pages", &pages);
 
         let mut sections = self
             .library
@@ -835,7 +832,6 @@ impl Site {
             }
         }
         sections.sort_by(|a, b| a.permalink.cmp(&b.permalink));
-        context.insert("sections", &sections);
 
         let mut taxonomies = vec![];
         for taxonomy in &self.taxonomies {
@@ -869,11 +865,44 @@ impl Site {
             taxonomies.push(terms);
         }
 
-        context.insert("taxonomies", &taxonomies);
-        context.insert("config", &self.config);
+        // Group all sitemap entries in one vector
+        let mut all_sitemap_entries = Vec::new();
+        all_sitemap_entries.append(&mut pages);
+        all_sitemap_entries.append(&mut sections);
+        for terms in taxonomies {
+            let mut terms = terms;
+            all_sitemap_entries.append(&mut terms);
+        }
 
+        // Count total number of sitemap entries to include in sitemap
+        let total_number = all_sitemap_entries.len();
+        let sitemap_limit = 30000;
+
+        if total_number < sitemap_limit {
+            // Create single sitemap
+            let mut context = Context::new();
+            context.insert("sitemap_entries", &all_sitemap_entries);
             let sitemap = &render_template("sitemap.xml", &self.tera, context, &self.config.theme)?;
+            create_file(&self.output_path.join("sitemap.xml"), sitemap)?;
+            return Ok(())
+        }
         
+        // Create multiple sitemaps (max 30000 urls each)
+        let mut sitemap_index = Vec::new();
+        for (i, chunk) in all_sitemap_entries.chunks(sitemap_limit).enumerate() {
+            let mut context = Context::new();
+            context.insert("sitemap_entries", &chunk);
+            let sitemap = &render_template("sitemap.xml", &self.tera, context, &self.config.theme)?;
+            let file_name = format!("sitemap{}.xml", i+1);
+            create_file(&self.output_path.join(&file_name), sitemap)?;
+            let mut sitemap_url:String = self.config.make_permalink(&file_name);
+            sitemap_url.pop(); // Remove trailing slash
+            sitemap_index.push(sitemap_url);
+        }
+        // Create main sitemap that reference numbered sitemaps
+        let mut main_context = Context::new();
+        main_context.insert("sitemaps", &sitemap_index);
+        let sitemap = &render_template("split_sitemap_index.xml", &self.tera, main_context, &self.config.theme)?;
         create_file(&self.output_path.join("sitemap.xml"), sitemap)?;  
         
         Ok(())

components/templates/src/builtins/sitemap.xml

@@ -1,22 +1,10 @@
 <urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
-    {% for page in pages %}
+    {% for sitemap_entry in sitemap_entries %}
         <url>
-            <loc>{{ page.permalink | safe }}</loc>
-            {% if page.date %}
-                <lastmod>{{ page.date }}</lastmod>
+            <loc>{{ sitemap_entry.permalink | safe }}</loc>
+            {% if sitemap_entry.date %}
+                <lastmod>{{ sitemap_entry.date }}</lastmod>
             {% endif %}
         </url>
     {% endfor %}
-    {% for section in sections %}
-        <url>
-            <loc>{{ section.permalink | safe }}</loc>
-        </url>
-    {% endfor %}
-    {% for taxonomy in taxonomies %}
-        {% for entry in taxonomy %}
-        <url>
-            <loc>{{ entry.permalink | safe }}</loc>
-        </url>
-        {% endfor %}
-    {% endfor %}
 </urlset>

components/templates/src/builtins/split_sitemap_index.xml

@@ -0,0 +1,7 @@
+<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+    {% for sitemap in sitemaps %}
+        <sitemap>
+            <loc>{{ sitemap }}</loc>
+        </sitemap>
+    {% endfor %}
+</sitemapindex>

components/templates/src/lib.rs

@@ -35,7 +35,8 @@ lazy_static! {
             ("__zola_builtins/rss.xml", include_str!("builtins/rss.xml")),
             ("__zola_builtins/sitemap.xml", include_str!("builtins/sitemap.xml")),
             ("__zola_builtins/robots.txt", include_str!("builtins/robots.txt")),
-            ("anchor-link.html", include_str!("builtins/anchor-link.html")),
+            ("__zola_builtins/split_sitemap_index.xml", include_str!("builtins/split_sitemap_index.xml")),
+            ("__zola_builtins/anchor-link.html", include_str!("builtins/anchor-link.html")),
             (
                 "__zola_builtins/shortcodes/youtube.html",
                 include_str!("builtins/shortcodes/youtube.html"),