From 7f75aed146b87e2e70e597dec70f289fedd31a70 Mon Sep 17 00:00:00 2001
From: "google-labs-jules[bot]"
 <161369871+google-labs-jules[bot]@users.noreply.github.com>
Date: Mon, 1 Dec 2025 14:31:54 +0000
Subject: [PATCH] feat: Add sitemap generator

Adds the `sitemap_generator` gem to generate a sitemap.

- Includes static routes and dynamic routes for active crops, plantings, seeds, photos, posts, and members.
- Implements a caching mechanism to ensure the sitemap is generated no more frequently than every 72 hours.
- Updates `robots.txt` to point to the new sitemap.
- Includes a unit test for the caching Rake task.
---
 Gemfile                           |  2 ++
 config/sitemap.rb                 | 50 +++++++++++++++++++++++++++++++
 lib/tasks/sitemap_cached.rake     | 19 ++++++++++++
 public/robots.txt                 |  2 ++
 spec/tasks/sitemap_cached_spec.rb | 38 +++++++++++++++++++++++
 5 files changed, 111 insertions(+)
 create mode 100644 config/sitemap.rb
 create mode 100644 lib/tasks/sitemap_cached.rake
 create mode 100644 spec/tasks/sitemap_cached_spec.rb

diff --git a/Gemfile b/Gemfile
index 8c85fe032..a633539a1 100644
--- a/Gemfile
+++ b/Gemfile
@@ -202,3 +202,5 @@ end
 gem "i18n_data", "~> 1.1"
 
 gem "paper_trail", "~> 17.0"
+
+gem 'sitemap_generator'
diff --git a/config/sitemap.rb b/config/sitemap.rb
new file mode 100644
index 000000000..7c13f356a
--- /dev/null
+++ b/config/sitemap.rb
@@ -0,0 +1,50 @@
+require 'rubygems'
+require 'sitemap_generator'
+
+SitemapGenerator::Sitemap.default_host = 'https://growstuff.org'
+SitemapGenerator::Sitemap.sitemaps_host = ENV['SITEMAP_HOST']
+SitemapGenerator::Sitemap.public_path = 'public/'
+SitemapGenerator::Sitemap.sitemaps_path = 'sitemaps/'
+
+SitemapGenerator::Sitemap.create do
+  # static pages
+  add '/community', changefreq: 'weekly', priority: 0.7
+  add '/terms', changefreq: 'yearly', priority: 0.2
+  add '/privacy', changefreq: 'yearly', priority: 0.2
+  add '/about', changefreq: 'monthly', priority: 0.4
+  add '/faq', changefreq: 'monthly', priority: 0.4
+  add '/contact', changefreq: 'yearly', priority: 0.2
+  add '/members', changefreq: 'weekly', priority: 0.6
+  add '/gardens', changefreq: 'weekly', priority: 0.6
+  add '/plantings', changefreq: 'weekly', priority: 0.6
+  add '/harvests', changefreq: 'weekly', priority: 0.6
+  add '/seeds', changefreq: 'weekly', priority: 0.6
+  add '/crops', changefreq: 'weekly', priority: 0.6
+  add '/posts', changefreq: 'daily', priority: 0.8
+  add '/forums', changefreq: 'daily', priority: 0.8
+  add '/photos', changefreq: 'daily', priority: 0.8
+
+  Crop.approved.find_each do |crop|
+    add crop_path(crop), lastmod: crop.updated_at
+  end
+
+  Planting.active.find_each do |planting|
+    add planting_path(planting), lastmod: planting.updated_at
+  end
+
+  Seed.active.find_each do |seed|
+    add seed_path(seed), lastmod: seed.updated_at
+  end
+
+  Photo.find_each do |photo|
+    add photo_path(photo), lastmod: photo.updated_at
+  end
+
+  Post.find_each do |post|
+    add post_path(post), lastmod: post.updated_at
+  end
+
+  Member.kept.find_each do |member|
+    add member_path(member), lastmod: member.updated_at
+  end
+end
diff --git a/lib/tasks/sitemap_cached.rake b/lib/tasks/sitemap_cached.rake
new file mode 100644
index 000000000..9d6b43073
--- /dev/null
+++ b/lib/tasks/sitemap_cached.rake
@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+namespace :sitemap do
+  desc 'Generate sitemap, but only if it has not been generated in the last 72 hours'
+  task cached_refresh: :environment do
+    sitemap_file = Rails.root.join('tmp', 'sitemap_generated_at.txt')
+    if File.exist?(sitemap_file)
+      last_generated_at = Time.parse(File.read(sitemap_file))
+      if last_generated_at > 72.hours.ago
+        puts 'Sitemap has been generated within the last 72 hours. Skipping.'
+        exit
+      end
+    end
+
+    Rake::Task['sitemap:refresh'].invoke
+
+    File.write(sitemap_file, Time.now.to_s)
+  end
+end
diff --git a/public/robots.txt b/public/robots.txt
index da250e302..d9421bb8c 100644
--- a/public/robots.txt
+++ b/public/robots.txt
@@ -140,3 +140,5 @@ Crawl-delay: 5
 #
 User-agent: *
 Disallow: /api/
+
+Sitemap: https://growstuff.org/sitemaps/sitemap.xml.gz
diff --git a/spec/tasks/sitemap_cached_spec.rb b/spec/tasks/sitemap_cached_spec.rb
new file mode 100644
index 000000000..1dbb0855a
--- /dev/null
+++ b/spec/tasks/sitemap_cached_spec.rb
@@ -0,0 +1,38 @@
+require 'rails_helper'
+require 'rake'
+
+describe 'sitemap:cached_refresh' do
+  let(:sitemap_file) { Rails.root.join('tmp', 'sitemap_generated_at.txt') }
+
+  before :all do
+    Rails.application.load_tasks
+  end
+
+  before :each do
+    Rake::Task['sitemap:refresh'].clear
+    File.delete(sitemap_file) if File.exist?(sitemap_file)
+  end
+
+  after :each do
+    File.delete(sitemap_file) if File.exist?(sitemap_file)
+  end
+
+  it 'calls sitemap:refresh if the cache file does not exist' do
+    expect(Rake::Task['sitemap:refresh']).to receive(:invoke)
+    Rake::Task['sitemap:cached_refresh'].invoke
+  end
+
+  it 'calls sitemap:refresh if the cache file is older than 72 hours' do
+    File.write(sitemap_file, 73.hours.ago.to_s)
+    expect(Rake::Task['sitemap:refresh']).to receive(:invoke)
+    Rake::Task['sitemap:cached_refresh'].reenable
+    Rake::Task['sitemap:cached_refresh'].invoke
+  end
+
+  it 'does not call sitemap:refresh if the cache file is newer than 72 hours' do
+    File.write(sitemap_file, 71.hours.ago.to_s)
+    expect(Rake::Task['sitemap:refresh']).not_to receive(:invoke)
+    Rake::Task['sitemap:cached_refresh'].reenable
+    Rake::Task['sitemap:cached_refresh'].invoke
+  end
+end