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_rendering.py

@@ -95,4 +95,58 @@ class TestApplyColormap:
         tif_file = _make_test_tif(tmp_path, data)
         result = tif_to_png(tif_file, tmp_path, 5.0)
         assert result is not None
-        assert result.exists()
+        assert result.exists()
+
+
+class TestTifToCrop:
+    """Conversion TIF → dalle cartographique (tif_to_crop)."""
+
+    @staticmethod
+    def _write_named_tif(tmp_path, name, arr):
+        transform = from_bounds(660000, 6700000, 661000, 6701000, arr.shape[1], arr.shape[0])
+        tif_file = tmp_path / name
+        with rasterio.open(
+            tif_file, 'w', driver='GTiff', height=arr.shape[0], width=arr.shape[1],
+            count=1, dtype='float32', crs='EPSG:2154', transform=transform,
+            nodata=float('nan'), compress='lzw'
+        ) as dst:
+            dst.write(arr.astype('float32'), 1)
+        return tif_file
+
+    def test_nodata_renders_black(self, tmp_path):
+        """Le nodata restant est rendu en noir (comportement historique).
+
+        Les trous du MNT sont comblés en amont (interpolation dans
+        create_dtm_fast, tous modes) ; ce qui reste en nodata doit rester
+        visible en noir sur la dalle plutôt qu'inventé au rendu.
+        """
+        from PIL import Image as PILImage
+        from lidar_pipeline.rendering import tif_to_crop
+
+        data = np.random.default_rng(7).normal(50, 10, (40, 40)).astype(np.float32)
+        data[15:25, 15:25] = np.nan
+        tif_file = self._write_named_tif(tmp_path, "LHD_test_slope.tif", data)
+
+        # WebP lossless : l'encodeur AVIF de l'image « saigne » légèrement les
+        # bords du noir même en lossless — on teste la logique nodata→noir,
+        # pas les artefacts du codec.
+        out = tif_to_crop(tif_file, tmp_path, 5.0, keep_tif=True,
+                          quality=100, output_format='webp')
+        assert out is not None and out.exists()
+
+        rgb = np.asarray(PILImage.open(str(out)).convert('RGB'))
+        hole = rgb[15:25, 15:25, :]
+        assert np.all(hole == 0), "le nodata doit être rendu en noir"
+
+    def test_without_nodata(self, tmp_path):
+        """Un TIF sans nodata est converti sans crash, taille préservée."""
+        from PIL import Image as PILImage
+        from lidar_pipeline.rendering import tif_to_crop
+
+        data = np.random.default_rng(7).normal(50, 10, (40, 40)).astype(np.float32)
+        tif_file = self._write_named_tif(tmp_path, "LHD_test_slope.tif", data)
+
+        out = tif_to_crop(tif_file, tmp_path, 5.0)
+        assert out is not None and out.exists()
+        img = PILImage.open(str(out))
+        assert img.size == (40, 40)