Images: add Debian auto-update script

It's parsing the webpages, it's very MEH. There's sadly no API/JSON to
figure out what the latest release is. We do have to parse that HTML…

debian/default.nix

@@ -2,8 +2,8 @@
 let
   imagesJSON = lib.importJSON ./images.json;
   fetchImage = image: pkgs.fetchurl {
-    inherit (image) hash;
-    url = "https://cloud.debian.org/images/cloud/${image.name}";
+    sha256 = image.hash;
+    url = image.name;
   };
   images = lib.mapAttrs (k: v: fetchImage v) imagesJSON.${system};
   makeVmTestForImage = image: { testScript, sharedDirs, diskSize ? null }: generic.makeVmTest {

debian/images.json

@@ -1,12 +1,22 @@
 {
+  "aarch64-linux": {
+    "12": {
+      "hash": "0sckj4cpa9knd4dqab2l6z79gms7p5jq983408dvrfmb7dpz7p54",
+      "name": "https://cloud.debian.org/images/cloud/bookworm/20240211-1654/debian-12-generic-arm64-20240211-1654.qcow2"
+    },
+    "13": {
+      "hash": "0l6lh2b54n9pkdc7b12qpj7qbw8ql84a0cf4c97ml22d9wxjmydk",
+      "name": "https://cloud.debian.org/images/cloud/trixie/daily/20240328-1700/debian-13-generic-arm64-daily-20240328-1700.qcow2"
+    }
+  },
   "x86_64-linux": {
     "12": {
-      "name": "bookworm/20240211-1654/debian-12-generic-amd64-20240211-1654.qcow2",
-      "hash": "sha256-M9B0CoAUtHvu4vvftqAlrmOrnAMH3pwX9Lr0ir9nB8M="
+      "hash": "1hq7cyzqmx5syhbrrph70ffanqxf4nhbdpzvwbp7pd0lh0579l1k",
+      "name": "https://cloud.debian.org/images/cloud/bookworm/20240211-1654/debian-12-generic-amd64-20240211-1654.qcow2"
     },
     "13": {
-      "name": "trixie/daily/20240220-1663/debian-13-generic-amd64-daily-20240220-1663.qcow2",
-      "hash": "sha256-JBoGFc68zgsiH6bZ9MmgXzBFAozfBct6Bajjnn08s3g="
+      "hash": "1p12msg0b2cnw9pp01n5dpzk2gc6pxm6y2m50dzzf0s8gwp4p7md",
+      "name": "https://cloud.debian.org/images/cloud/trixie/daily/20240328-1700/debian-13-generic-amd64-daily-20240328-1700.qcow2"
     }
   }
 }

scripts/update-images.py

@@ -0,0 +1,72 @@
+#! /usr/bin/env nix-shell
+#! nix-shell -i python3 -p python3 python3Packages.beautifulsoup4 python3Packages.requests nix-prefetch
+
+import re
+import requests
+import subprocess
+import json
+from bs4 import BeautifulSoup
+from datetime import datetime
+
+
+def nix_hash(url):
+    print(f"[+] Calculating Nix hash for {url}")
+    res = subprocess.run(["nix-prefetch-url", url], stdout=subprocess.PIPE)
+    return res.stdout.rstrip().decode("utf-8")
+
+def get_latest_debian_image(url):
+    print(f"[+] Parsing debian index {url}")
+    # Step 1: retrieve the latest entry
+    page = requests.get(url)
+    soup = BeautifulSoup(page.content, "html.parser")
+    rows = soup.find_all("tr")
+    l = [row.a["href"] for row in rows if row.a]
+    # Filtering out non-datetime entries such as "daily" or "latest"
+    l = [s for s in l if re.compile("^[0-9]{8}-[0-9]{4}/$").match(s)]
+    # Parsing date part of the string
+    parsed_l = [(datetime.strptime(s[:8], '%Y%m%d'), s) for s in l]
+    latest = max(parsed_l)
+    url = f"{url}/{latest[1]}"
+    print(f"[+] Parsing latest entry: {url}")
+
+    # Step 2: parse entry
+    page = requests.get(url)
+    soup = BeautifulSoup(page.content, "html.parser")
+    rows = soup.find_all("tr")
+    l = [row.a["href"] for row in rows if row.a]
+    res = {}
+    for s in l:
+        if re.compile("^.*-generic-.*\.qcow2$").match(s):
+            if "amd64" in s:
+                res["x86_64-linux"] = f"{url}{s}"
+            elif "arm64" in s:
+                res["aarch64-linux"] = f"{url}{s}"
+    return res
+
+"""
+Parse the debian cloudimages
+"""
+def debian_parse():
+    bookworm_url = "https://cloud.debian.org/images/cloud/bookworm"
+    trixie_url = "https://cloud.debian.org/images/cloud/trixie/daily"
+    bookworm = get_latest_debian_image(bookworm_url)
+    trixie = get_latest_debian_image(trixie_url)
+    res = {}
+    for arch in bookworm.keys():
+        res[arch] = {
+            "12": {
+                "name": bookworm[arch],
+                "hash": nix_hash(bookworm[arch])
+
+            },
+            "13": {
+                "name": trixie[arch],
+                "hash": nix_hash(trixie[arch])
+            }
+        }
+    return json.dumps(res)
+
+if __name__ == '__main__':
+    debian_json = debian_parse()
+    with open("debian.json", "w") as f:
+        f.write(debian_json)