Test diagonal of inductance matrix

Use different number of quadrature points and two different implementations

import numpy as np
import matplotlib.pyplot as plt

from bfieldtools.utils import load_example_mesh

coilmesh = load_example_mesh("10x10_plane")

from bfieldtools.mesh_impedance import self_inductance_matrix, mutual_inductance_matrix


M1 = mutual_inductance_matrix(coilmesh, coilmesh, quad_degree=1)
M2 = mutual_inductance_matrix(coilmesh, coilmesh, quad_degree=2)
M3 = mutual_inductance_matrix(coilmesh, coilmesh, quad_degree=3)

plt.figure()
for m in (M1, M2, M3):
    plt.plot(np.diag(m))

MM1 = self_inductance_matrix(coilmesh, quad_degree=1, analytic_self_coupling=True)
MM2 = self_inductance_matrix(coilmesh, quad_degree=2, analytic_self_coupling=True)
MM3 = self_inductance_matrix(coilmesh, quad_degree=3, analytic_self_coupling=True)

Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (14013 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix
Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (14045 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix
Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (13976 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix
Computing self-inductance matrix using rough quadrature (degree=1).              For higher accuracy, set quad_degree to 4 or more.
Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (14004 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix
Computing self-inductance matrix using rough quadrature (degree=2).              For higher accuracy, set quad_degree to 4 or more.
Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (13980 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix
Estimating 2432 MiB required for 676 by 676 vertices...
Computing inductance matrix in 20 chunks (13968 MiB memory free),              when approx_far=True using more chunks is faster...
Computing triangle-coupling matrix

plt.gca().set_prop_cycle(None)
for m in (MM1, MM2, MM3):
    plt.plot(np.diag(m), "--")