Simplify line plots along cross-section

[veenstrajelmer]

When slicing a dataset, it is very complex to get the variables without depth component (like bedlevel, waterlevel) out of it. An example for bedlevel is given below. The desired result is the bed level along the slice line_array in a line plot with distance along the line_array on the xaxis. This should be simplified.

import matplotlib.pyplot as plt
plt.close('all')
import numpy as np
import dfm_tools as dfmt

file_nc = dfmt.data.fm_grevelingen_map(return_filepath=True)
uds = dfmt.open_partitioned_dataset(file_nc)
line_array = np.array([[ 53181.96942503, 424270.83361629],
                       [ 55160.15232593, 416913.77136685]])

# fig, ax = plt.subplots()
# uds['mesh2d_flowelem_bl'].ugrid.plot(cmap='jet') #TODO: default is edgecolor='face', should work even better with edgecolor='none', but that results in seethrough edges anyway, report to matplotlib?
# ax.plot(line_array[:,0],line_array[:,1],'b',linewidth=3)

xr_crs_ugrid = dfmt.polyline_mapslice(uds.isel(time=0).max(dim="nmesh2d_layer").max(dim="nmesh2d_interface"), line_array)
fig, ax = plt.subplots()
xr_crs_ugrid['mesh2d_flowelem_bl'].ugrid.plot()

fig, ax = plt.subplots()
xr_crs_ugrid['mesh2d_flowelem_bl'].plot()

# s_array = xr_crs_ugrid.grid.face_coordinates[:,0] #contains nans for some reason
s_array_nonan = xr_crs_ugrid.grid.node_coordinates[:,0][xr_crs_ugrid.grid.face_node_connectivity].mean(axis=1)
bed_array = xr_crs_ugrid['mesh2d_flowelem_bl'].to_numpy()
fig, ax = plt.subplots()
ax.plot(s_array_nonan,bed_array)

Looks nice but not desired:

Desired, but complex workaround needed at this moment:

A simpler alternative, but results in degrees on x-axis for spherical models:

from shapely import LineString
uds_crs = uds.ugrid.intersect_linestring(LineString(line_array))
fig,ax = plt.subplots()
uds_crs.mesh2d_flowelem_bl.plot(x="mesh2d_s")