Fix interpolation and spline issues (#161)

pyproject.toml

@@ -44,16 +44,16 @@ dependencies = [
 # https://peps.python.org/pep-0621/#dependencies-optional-dependencies
 [project.optional-dependencies]
 all = [
-    "napari[all]>=0.4.19rc2",
+    "napari[all]>=0.4.19rc4",
     "napari-properties-plotter",
     "napari-properties-viewer",
-    "napari-label-interpolator",
+    "napari-label-interpolator>=0.1.1",
 ]
 test = [
     "pytest>=6.0",
     "pytest-cov",
     "pytest-qt",
-    "napari[all]>=0.4.19rc2",
+    "napari[all]>=0.4.19rc4",
 ]
 dev = [
     "blik[test]",

src/blik/widgets/picking.py

@@ -42,6 +42,9 @@ def _generate_surface_grids_from_shapes_layer(
             for line in sorted(lines, key=lambda x: x[0, 0])
         ]
 
+        # drop duplicate points (messes up scipy's fitpack for splines)
+        lines = [pd.DataFrame(line).drop_duplicates().to_numpy() for line in lines]
+
         try:
             surface_grids.append(
                 GriddedSplineSurface(
@@ -82,7 +85,10 @@ def _resample_surfaces(image_layer, surface_grids, spacing, thickness, masked):
 def _generate_filaments_from_points_layer(filament_picks):
     """create a new filament representation from picked points."""
     # invert xyz to go back to xyz world
-    return HelicalFilament(points=invert_xyz(filament_picks.data.astype(float)))
+    points = invert_xyz(filament_picks.data.astype(float))
+    # drop duplicate points (messes up scipy's fitpack for splines)
+    points = pd.DataFrame(points).drop_duplicates().to_numpy()
+    return HelicalFilament(points=points)
 
 
 def _resample_filament(image_layer, filament, spacing, thickness):