Main allocation network

[SouthEndMusic]

Fixes #768.

[SouthEndMusic]

@visr do you think something like this is useful as a plotting option?

from ribasim_testmodels import subnetwork_model
import matplotlib.pyplot as plt
import numpy as np
from scipy.spatial import ConvexHull

model = subnetwork_model()
df = model.network.node.df

for allocation_subnetwork_id, df_subnetwork in df.groupby("allocation_network_id"):
    points = np.stack((df.geometry.x.to_numpy(), df.geometry.y.to_numpy()), axis=1)
    hull = ConvexHull(points)
    bounding_polygon_vertices = points[hull.vertices, :]
    hull_center = np.mean(bounding_polygon_vertices, axis=0)
    bounding_polygon_vertices = (
        bounding_polygon_vertices - hull_center
    ) * 1.1 + hull_center

    ax = model.plot()
    ax.fill(
        bounding_polygon_vertices[:, 0],
        bounding_polygon_vertices[:, 1],
        linestyle="--",
        facecolor="none",
        linewidth=2,
        edgecolor="gray",
    )
    ax.text(
        *hull_center,
        f"Subnetwork {allocation_subnetwork_id}",
        horizontalalignment="center",
    )

plt.show()

P.s.: I am aware that hull_center is not correct, but I didn't find a centroid function in the current Python environment.

[SouthEndMusic]

This PR is superseded by among others #914.