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

lidar_pipeline/tests/test_webapp.py

@@ -0,0 +1,81 @@
+"""Tests du serveur web de génération de zones (webapp)."""
+
+
+def test_bbox_to_cells_single_km_cell():
+    """Une bbox couvrant ~1 km² retourne la cellule L93 correspondante."""
+    from lidar_pipeline.webapp import bbox_to_cells
+    # Cellule 1054,6882 : X∈[1054000,1055000], Y∈[6881000,6882000] (L93)
+    from rasterio.warp import transform as warp_transform
+    lons, lats = warp_transform('EPSG:2154', 'EPSG:4326',
+                                [1054100, 1054900], [6881100, 6881900])
+    cells = bbox_to_cells(min(lons), min(lats), max(lons), max(lats))
+    assert (1054, 6882) in cells
+    # La sélection reste locale : pas de cellule lointaine
+    for (c, r) in cells:
+        assert abs(c - 1054) <= 1 and abs(r - 6882) <= 1
+
+
+def test_bbox_to_cells_empty_for_tiny_bbox():
+    """Une bbox quasi ponctuelle ne sélectionne rien (rétrécie sous 1 m)."""
+    from lidar_pipeline.webapp import bbox_to_cells
+    assert bbox_to_cells(7.850000, 48.930000, 7.850001, 48.930001) == []
+
+
+def test_processed_cells(tmp_path):
+    """processed_cells lit les dossiers de visualisations."""
+    from lidar_pipeline.webapp import processed_cells
+    vis = tmp_path / "visualisations" / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69_r0p2"
+    vis.mkdir(parents=True)
+    (vis / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69_r0p2_aspect.avif").write_bytes(b"x")
+    assert processed_cells(tmp_path) == {(1054, 6882)}
+
+
+def test_missing_cells_filters_processed(tmp_path):
+    """Les cellules déjà traitées sont exclues, les autres gardent leurs coins."""
+    from lidar_pipeline.webapp import missing_cells_with_corners
+    vis = tmp_path / "visualisations" / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69"
+    vis.mkdir(parents=True)
+    (vis / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69_aspect.avif").write_bytes(b"x")
+    todo = missing_cells_with_corners([(1054, 6882), (1055, 6882)], tmp_path)
+    assert len(todo) == 1
+    assert todo[0]['col'] == 1055 and todo[0]['row'] == 6882
+    assert len(todo[0]['corners']) == 4  # SW, SE, NE, NW
+
+
+def test_missing_cells_include_done(tmp_path):
+    """include_done=True conserve les cellules déjà traitées (régénération)."""
+    from lidar_pipeline.webapp import missing_cells_with_corners
+    vis = tmp_path / "visualisations" / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69"
+    vis.mkdir(parents=True)
+    (vis / "LHD_FXX_1054_6882_PTS_LAMB93_IGN69_aspect.avif").write_bytes(b"x")
+    todo = missing_cells_with_corners([(1054, 6882), (1055, 6882)], tmp_path,
+                                      include_done=True)
+    assert {(t['col'], t['row']) for t in todo} == {(1054, 6882), (1055, 6882)}
+
+
+def test_build_command_regenerate():
+    """regenerate=True ajoute --force --force-classification à la commande."""
+    from lidar_pipeline.webapp import _build_command
+    cmd = " ".join(_build_command([(1054, 6882)], regenerate=True))
+    assert "--force" in cmd
+    assert "--force-classification" in cmd
+    cmd = " ".join(_build_command([(1054, 6882)]))
+    assert "--force" not in cmd
+    assert "--force-classification" not in cmd
+
+
+def test_build_command_ground_classification():
+    """La commande utilise la méthode de classification demandée (défaut : ign)."""
+    from lidar_pipeline.webapp import _build_command, GROUND_CLASS_METHODS
+    # Défaut : ign (pré-classification)
+    cmd = _build_command([(1054, 6882)])
+    i = cmd.index("--ground-classification")
+    assert cmd[i + 1] == "ign"
+    # Chaque méthode valide est transmise telle quelle, avec ou sans régénération
+    for method in GROUND_CLASS_METHODS:
+        for regenerate in (False, True):
+            cmd = _build_command([(1054, 6882)], regenerate=regenerate, ground_class=method)
+            i = cmd.index("--ground-classification")
+            assert cmd[i + 1] == method
+            assert ("--force" in cmd) == regenerate
+            assert ("--force-classification" in cmd) == regenerate