Advice on plotting

[mohawk811]

Hello!

I was wondering if I could get some input on some custom plots I'm trying to do. For this code I want to plot QS error on the Y axis and aspect ratio on the X axis. I expect to see that as I go to higher aspect ratio QS error should decrease but my plot is showing otherwise. Could I get some advice or input on how to show this?

import matplotlib.pyplot as plt
plt.rcParams.update({'font.size': 22})
import numpy as np

# Lists to store aspect ratios, elongations, QS errors, and corresponding 'a' values
aspect_ratios = []
elongations = []
qs_errors = []
file_aspect = []

# Loop through each equilibrium
for a in range(4, 20):
    # Generate the filename based on the current elongation value 'a'
    filename = f'eq_opt_qs_3_aspect_{a}_10_29.h5'

    # Load the equilibrium object
    eq = desc.io.load(filename)

    # Define the grid
    grid = LinearGrid(M=eq.M_grid, N=eq.N_grid, NFP=eq.NFP, rho=np.array([.2,.6,1.0]), sym=True)

    # Compute the aspect ratio R0/a
    aspect_ratio = eq.compute('R0/a', grid=grid)['R0/a'].max()

    # Compute the elongation a_major/a_minor
    elongation_value = eq.compute('a_major/a_minor', grid=grid)['a_major/a_minor'].max()

    # Compute the QS error
    qs_error = eq.compute('f_C', grid=grid, helicity=(0,1))['f_C'].max()  # Assuming 'f_C' is the QS error

    # Append the values to the lists
    aspect_ratios.append(aspect_ratio)
    elongations.append(elongation_value)
    qs_errors.append(qs_error)
    file_aspect.append(a)  # Store the value of 'a' from the filename

# Plotting Aspect Ratio vs. Elongation with QS Error Color Scale
plt.figure(figsize=(10, 6))
sc = plt.scatter(aspect_ratios, elongations, c=qs_errors, cmap='viridis', marker='o', vmin=0, vmax=.5)

for i in range(len(aspect_ratios)):
    # Slight offset for the text position
    x_offset = 0.02 * (i % 2 - 0.5)
    y_offset = 0.02 * (i % 2 - 0.5)

    # Aspect ratio and elongation annotation
    plt.text(aspect_ratios[i] + x_offset, elongations[i] + y_offset,
             f'({aspect_ratios[i]:.2f}, {elongations[i]:.2f})',
             fontsize=9, ha='right', va='bottom', bbox=dict(facecolor='white', alpha=0.5))

    # Annotate with the 'a' value from the filename in red with a slight offset
    plt.text(aspect_ratios[i] - x_offset, elongations[i] - y_offset,
             f'a={file_aspect[i]:.2f}', fontsize=9, ha='left', va='top', color='red',
             bbox=dict(facecolor='white', alpha=0.5))

plt.xlabel('Aspect Ratio (R0/a)')
plt.ylabel('Average Elongation (a_major/a_minor)')
plt.title('Aspect Ratio vs. Average Elongation with QS Error Color Scale')
plt.colorbar(sc, label='QS Error (f_C)')  # Add a colorbar to represent QS error
plt.grid(True)
plt.show()

# New plot: Aspect Ratio vs. QS Error (f_C)
plt.figure(figsize=(10, 6))
plt.plot(aspect_ratios, qs_errors, marker='o', color='b', label='QS Error (f_C)')

# Annotating 'a' values
for i in range(len(aspect_ratios)):
    plt.text(aspect_ratios[i], qs_errors[i], f'a={file_aspect[i]}', fontsize=9, ha='right', va='bottom', color='red')

plt.xlabel('Aspect Ratio (R0/a)')
plt.ylabel('QS Error (f_C)')
plt.title('Aspect Ratio vs. QS Error')
plt.grid(True)
plt.legend()
plt.show()

[mohawk811]

Here are the generated plots:

[mohawk811]

[dpanici]

hard to say much without knowing the specifics of the optimization and how good of QP solutions the individual equilibria are, but some ideas for the plots

• QP is impossible to get near-axis, but I see your grid has a point at rho=0.2, perhaps re-plotting with a grid just on the boundary would show different results? if the QP error increase near-axis for a given equilibrium, that may explain this
• if you plot the QS error with say plot_qs_error and compare the highest and lowest AR, are they similar/what you expect (though this function does normalize the error metrics so that may be hard to draw conclusions from)
• if you plot_boozer_surface for the highest AR versus the lowest AR, is it obvious one is better than the other?