Add web map with zone generation API, side job queue, and restore historical DTM hole rendering

- webapp.py: FastAPI serving the continuous map (port 8973) with
  /api/preview, /api/generate and /api/status; tiles are downloaded
  from IGN and processed in a logged subprocess, tracked live in a
  side "File de génération" panel that survives page reloads
- fetch_ign.py: download missing 1 km LiDAR HD tiles from the IGN
  geoplateforme before processing
- index.py: tile thumbnails and 500 m subtiles are now invalidated by
  mtime so regenerating a tile refreshes its cached images; progress
  logging per tile
- dtm.py: back to the historical gap handling (small gaps filled by
  fillnodata only, larger holes left as nodata rendered black);
  lowest-return floor only via --bare-earth, IGN class selection via
  --ign-classes
- cli.py: positional input now optional (--rebuild-index works alone)
- docker-compose.yml: serve (GPU, port 8973) and process services;
  launch via docker compose only (documented in AGENTS.md/AGENTS.md)
- tests: 131 passing, incl. regressions for thumbnail staleness,
  --rebuild-index without input, and nodata rendering

docker-compose.yml

@@ -1,38 +1,61 @@
-version: '3.8'
+# Lancement du pipeline LiDAR — TOUJOURS via docker compose :
+#   docker compose build            # après chaque édition de code (code baké dans l'image)
+#   docker compose up -d serve      # carte interactive + API sur http://localhost:8973
+#   docker compose logs -f serve    # journal du serveur
+#   docker compose down             # arrêt
+# Traitement ponctuel (sans serveur) :
+#   docker compose run --rm process [-r 0.5,0.2 | --force | --file ...]
+# Jupyter (opt-in) :
+#   docker compose --profile interactive up -d jupyter
 
 services:
-  lidar:
+  # Carte interactive + API de génération (mode ./run.sh --serve)
+  serve:
     build: .
-    container_name: lidar-archeo
+    image: lidar-lidar
+    container_name: lidar-serve
+    init: true
     user: "1000:1000"
+    gpus: all
+    ports:
+      - "8973:8973"
     volumes:
-      # Mount your LAZ files directory here
-      - ./input:/data/input:ro
-      # Output directory
+      # input/ en écriture : l'API y télécharge les dalles IGN manquantes
+      - ./input:/data/input
       - ./output:/data/output
-      # Optional: Mount a large data directory
-      # - /path/to/your/laz/files:/data/input:ro
     environment:
       - TZ=Europe/Paris
-      # Processing parameters
-      - RESOLUTION=0.5
-      - WHITEBOX_THREADS=4
-    # Resource limits (adjust based on your system)
-    deploy:
-      resources:
-        limits:
-          cpus: '4'
-          memory: 8G
-        reservations:
-          cpus: '2'
-          memory: 4G
-    # Override default command
-    command: ["process_lidar.py", "/data/input", "-o", "/data/output", "-r", "0.5"]
+      - LIDAR_INPUT_DIR=/data/input
+      - LIDAR_OUTPUT_DIR=/data/output
+      # Les générations lancées depuis la carte utilisent le GPU
+      - LIDAR_GPU=1
+      - LIDAR_WORKERS=2
+    command: python3 -m uvicorn lidar_pipeline.webapp:app --host 0.0.0.0 --port 8973
+    restart: unless-stopped
 
-  # Optional: Jupyter notebook for interactive exploration
+  # Traitement ponctuel des dalles input/ (une passe puis arrêt)
+  process:
+    build: .
+    image: lidar-lidar
+    container_name: lidar-process
+    init: true
+    user: "1000:1000"
+    gpus: all
+    volumes:
+      - ./input:/data/input
+      - ./output:/data/output
+    environment:
+      - TZ=Europe/Paris
+    command: ["python3", "-m", "lidar_pipeline", "/data/input", "-o", "/data/output", "-r", "0.5,0.2", "-g", "all"]
+    profiles:
+      - process
+
+  # Exploration interactive (opt-in : --profile interactive)
   jupyter:
     build: .
+    image: lidar-lidar
     container_name: lidar-jupyter
+    init: true
     ports:
       - "8888:8888"
     volumes: