Improve "fringy" look when plotting shaded dfsu.

[zweihuehner]

Description
When plotting a dfsu DataArray with plot_type="shaded", the triangles that lie close to elements that fell "dry" during a simulation are looking quite fringy. The underlying function is the matplotlib tripcolor in _FM_utils.py. Possibly the behavior is coming from matplotlib and can not be changed in mikeio.

To Reproduce
Load a dfsu dataset that comes from a m3fm simulation with dry falling elements. Plot a DataArray from the dataset like the following.

import mikeio
import matplotlib.pyplot as plt
ds = mikeio.read("example.dfsu")
fig, ax = plt.subplots(figsize = (14,14))
ds.Current_speed.plot(ax = ax, plot_type="shaded")

Example File:
example_dfsu.zip

Expected behavior
See a shaded plot without the fringy look of the triangle edges.

Screenshots

System information:

• Python version: 3.10.10
• MIKE IO version: 2.2.dev3

[ecomodeller]

It seems like support for plots for data with missing values, e.g. because of drying is generally lacking in MIKE IO.

In some cases it could be ok, to use the recently implemented mikeio.DataArray.fillna.

But in general filling missing values with a constant value, is very crude.

[ecomodeller]

Or how about removing all dry elements:

import matplotlib.pyplot as plt
import numpy as np
da = ds.Current_speed.isel(time=0)
fig, ax = plt.subplots(ncols=2, figsize=(15,4))
wet = ~np.isnan(da.to_numpy())
idx = np.where(wet)[0]
da.plot(plot_type="shaded", title="Default",ax=ax[0])

da.plot.outline(ax=ax[1])
da.isel(element=idx).plot(plot_type="shaded", ax=ax[1], title="Only wet elements");