Add serve-webapp.sh to manage the webapp stack on the lightweight machine

Single command to start the interactive map container in the background
(auto-restart after reboot), with stop/restart/status/sync/logs/update
subcommands. Configuration lives in webapp.env (template provided,
git-ignored): remote generation URL, shared token, rsync cache command,
periodic sync and allowed network. The SSH key is mounted automatically
for rsync, port conflicts are detected before launch, and update pulls the
repo, rebuilds the image and restarts.

Deploy doc updated: the script is now the recommended install path on the
Raspberry Pi, alongside run.sh (foreground) and docker compose.

AGENTS.md

@@ -4,6 +4,7 @@
 - build: `docker build -t lidar-lidar .`
 - build webapp légère (Raspberry Pi, déploiement 2 machines — cf. `docs/DEPLOY_WEBAPP.md`): `docker compose -f docker-compose.webapp.yml up -d --build` (image `Dockerfile.webapp`, sans PDAL/GPU)
 - build générateur de tuiles (machine de traitement): `docker compose -f docker-compose.worker.yml up -d --build` (service `worker`, API pour les webapp distantes)
+- stack webapp (machine légère): `./serve-webapp.sh [start|stop|restart|status|sync|logs|update]`, config dans `webapp.env` (modèle `webapp.env.example`, ignoré par git)
 - test all: `./run.sh --test` (rebuild automatique de l'image avant les tests ; en `docker run` direct, rebuild manuellement d'abord)
 - test file: `docker run --rm lidar-lidar python3 -m pytest -v --pyargs lidar_pipeline.tests.<module>`
 - test case: `docker run --rm lidar-lidar python3 -m pytest -v --pyargs lidar_pipeline.tests.<module>::<TestClass>::<test_method>`