Duplicate node in subsetting

[SorooshMani-NOAA]

The mesh in the seam section seem to sometimes include two nodes that are almost at the same spot. See https://github.com/noaa-ocs-modeling/SurgeTeamCoordination/issues/143#issuecomment-1844362533

And example in a local test is:
File seam_mesh.2dm

ND 7855 -7.7202011351507991E+01 3.5937568127835952E+01 0.0000000000000000E+00
ND 7856 -7.7202011351507991E+01 3.5937568127835959E+01 0.0000000000000000E+00

which then reappears in the final mesh as:

ND 1897211 -7.7202011351507991E+01 3.5937568127835959E+01 2.1569783265873127E+01
ND 1897212 -7.7202011351507991E+01 3.5937568127835952E+01 2.1569783265877248E+01

[SorooshMani-NOAA]

This issues seems to be coming from the geometry (shape) used for meshing the seam region (image in UTM17):

[SorooshMani-NOAA]

This issue is a results of this section of the code:

OCSMesh/ocsmesh/cli/subset_n_combine.py

Lines 630 to 633 in 0563a79

	poly_seam_3 = self._add_one_mesh_layer_to_polygon(
	poly_seam_2,
	mesh_fine.msh_t, poly_clip_hires_0,
	mesh_coarse.msh_t, poly_clip_lowres_0)

Here one layer from each mesh (hi and low-res) are taken - in addition to the clipped section by the clipper polygon and then added to the last iteration seam polygon (which doesn't have duplicate). This union operation results in the duplicate in this case:

OCSMesh/ocsmesh/cli/subset_n_combine.py

Lines 226 to 227 in 0563a79

	return unary_union(
	[polygon, poly_1_lyr_hires, poly_1_lyr_lowres])

poly_seam_2 doesn't have duplicate. Tested by debugging and using helper function:

def debug(poly):
    from shapely import get_coordinates, points
    gdf = gpd.GeoDataFrame(geometry=[poly]).explode()
    gdf = gdf[gdf.intersects(points(
                [842651.955858, 3.983695e+06] # Point in UTM17 from duplicate issue
    ).buffer(10))]
    if len(gdf) == 0:
        return
    return get_coordinates(gdf)

[SorooshMani-NOAA]

The issue seems to go away after dropping extra verts:

OCSMesh/ocsmesh/cli/subset_n_combine.py

Line 644 in 0563a79

poly_seam_7 = utils.drop_extra_vertex_from_polygon(poly_seam_6)

but reappears in poly_seam_8 (the final version) again due to operation:

OCSMesh/ocsmesh/cli/subset_n_combine.py

Lines 660 to 661 in 0563a79

	poly_seam_8 = poly_seam_7.difference(
	unary_union([poly_clip_hires, poly_clip_lowres]))

correction
Actually upon inspecting the resulting shape 8, it seems that those point do not reappear! So why do we have them in the seam mesh?

[SorooshMani-NOAA]

After the number of layers fix, if I use 8 layers, I do not get the issue with duplicate nodes, but with 9 I do!! WHY?!