From f59dc14d0432784f80e02c5062ac707a670e7f07 Mon Sep 17 00:00:00 2001
From: Daniel Micay <daniel.micay@grapheneos.org>
Date: Tue, 11 Jun 2024 11:45:12 -0400
Subject: [PATCH] generate sitemap as part of deployment and automate IndexNow

This allows adding the lastmod field based on the last modification date
of the file along with automating submission to IndexNow.
---
 deploy-static      | 10 +++++++++
 generate-sitemap   | 54 ++++++++++++++++++++++++++++++++++++++++++++++
 static/sitemap.xml | 45 --------------------------------------
 3 files changed, 64 insertions(+), 45 deletions(-)
 create mode 100755 generate-sitemap
 delete mode 100644 static/sitemap.xml

diff --git a/deploy-static b/deploy-static
index ce664878..9f30482e 100755
--- a/deploy-static
+++ b/deploy-static
@@ -16,10 +16,16 @@ remote=root@attestation.app
 
 # use last modified timestamps from attestation.app
 rsync -rptcv --chmod=D755,F644 --delete --fsync --preallocate $remote:/srv/attestation.app/ static-production
+rsync -pcv --chmod=D755,F644 --fsync --preallocate static-production/sitemap.xml{,.gz,.br} static-tmp/
 rsync -rpcv --chmod=D755,F644 --delete --fsync --preallocate static-tmp/ static-production
 for f in static-production/**.*(br|gz); do
     touch -r "${f%.*}" "$f"
 done
+changed="$(./generate-sitemap)"
+xmllint --noblanks static-tmp/sitemap.xml --output static-tmp/sitemap.xml
+brotli -f static-tmp/sitemap.xml
+zopfli static-tmp/sitemap.xml
+rsync -pcv --chmod=D755,F644 --fsync --preallocate static-tmp/sitemap.xml{,.gz,.br} static-production/
 
 active=$(ssh $remote readlink /srv/attestation.app)
 
@@ -48,3 +54,7 @@ rsync -pcv --chmod=755 --fsync --preallocate remote-backup $remote:/usr/local/bi
 rsync -pcv --chmod=644 --fsync --preallocate systemd/system/remote-backup.timer $remote:/etc/systemd/system/remote-backup.timer
 rsync -pcv --chmod=644 --fsync --preallocate systemd/system/remote-backup.service $remote:/etc/systemd/system/remote-backup.service
 rsync -pcv --chmod=644 --chown attestation:attestation --fsync --preallocate backup-public-key.txt cloud-archive.sh $remote:/var/lib/attestation/
+
+if [[ -n "$changed" ]]; then
+    ./indexnow <<< "$changed"
+fi
diff --git a/generate-sitemap b/generate-sitemap
new file mode 100755
index 00000000..17e51cd5
--- /dev/null
+++ b/generate-sitemap
@@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+
+from datetime import datetime, timezone
+from os.path import getmtime
+from pathlib import Path
+
+base = "https://attestation.app"
+
+pages = [
+    ["/", 1.0],
+    ["/.well-known/security.txt", 0.0],
+    ["/LICENSE.txt", 0.0],
+    ["/about", 1.0],
+    ["/donate", 0.5],
+    ["/contact", 0.1],
+    ["/humans.txt", 0.0],
+    ["/privacy-policy", 0.2],
+    ["/source", 0.1],
+    ["/tutorial", 1.0]
+]
+
+base_mtime = getmtime("static-tmp")
+entries = []
+
+for page in pages:
+    path = page[0]
+    loc = base + path
+    filepath = "static-production" + path
+    if path[-1] == '/':
+        filepath += "index.html"
+    elif "." not in path:
+        filepath += ".html"
+
+    mtime = getmtime(filepath)
+    if mtime > base_mtime:
+        print(loc)
+    lastmod = datetime.fromtimestamp(mtime, timezone.utc).strftime("%Y-%m-%dT%H:%M:%S%:z")
+    priority = page[1]
+    entries.append(f"""
+    <url>
+        <loc>{loc}</loc>
+        <lastmod>{lastmod}</lastmod>
+        <priority>{priority}</priority>
+    </url>""")
+
+sitemap = f"""<?xml version="1.0" encoding="UTF-8"?>
+<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
+        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+        xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd">{"".join(entries)}
+</urlset>
+"""
+
+with open("static-tmp/sitemap.xml", "w") as f:
+    f.write(sitemap)
diff --git a/static/sitemap.xml b/static/sitemap.xml
deleted file mode 100644
index e1bba6dc..00000000
--- a/static/sitemap.xml
+++ /dev/null
@@ -1,45 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
-        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
-        xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd">
-    <url>
-        <loc>https://attestation.app/</loc>
-        <priority>1.0</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/.well-known/security.txt</loc>
-        <priority>0.0</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/LICENSE.txt</loc>
-        <priority>0.0</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/about</loc>
-        <priority>1.0</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/donate</loc>
-        <priority>0.5</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/contact</loc>
-        <priority>0.1</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/humans.txt</loc>
-        <priority>0.0</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/privacy-policy</loc>
-        <priority>0.2</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/source</loc>
-        <priority>0.1</priority>
-    </url>
-    <url>
-        <loc>https://attestation.app/tutorial</loc>
-        <priority>1.0</priority>
-    </url>
-</urlset>