The coronal solution looks weird

[MohamedNedal]

I found that the field lines of the coronal solution for that day April 3, 2019, seem to be like weird hyperbola shapes in matplotlib. Since matplotlib doesn't show the footpoints of the field lines on the Sun, I tried to plot with pyvista but it seems to be very slow, and also the field lines look collimated around the polar axis of the Sun. Also in pyvista, the axes aren't shown although it's written in the code to show them.

Is there a way to fix the field lines?
Also is there a way to show the closed field lines in black, as in the pfsspy?

Here is the matplotlib code

mas_output = MASOutput(directory) # location of the predsci/MAS files 
br = mas_output['br']

# Do the tracing 
nseeds = 12
rad_extend = 1.5
r = np.ones(nseeds**2)*rad_extend*u.R_sun
lat = np.linspace(-89, 89, nseeds**2, endpoint=False)*u.deg
lon = np.random.rand(nseeds**2)*360*u.deg

max_steps = 10000
step_size = 1
tracer = FortranTracer(max_steps=max_steps, step_size=step_size)
flines = tracer.trace(mas_output, r=r, lat=lat, lon=lon)

fig = plt.figure(figsize=(8,8)) # figsize=(10,10) # figsize=plt.figaspect(1)
ax = plt.subplot(111, projection='3d')

# draw the Sun 
solrad = 1
usph, vsph = np.mgrid[0:2*np.pi:50j, 0:np.pi:50j]
xsph = np.cos(usph)*np.sin(vsph)*solrad
ysph = np.sin(usph)*np.sin(vsph)*solrad
zsph = np.cos(vsph)*solrad
ax.plot_surface(xsph, ysph, zsph, color='orange', alpha=1)

# draw the field lines 
idx = 0
for fline in flines:
    color = (br.sample_at_coords(np.mod(fline.lon[idx], 2*np.pi*u.rad), fline.lat[idx], fline.r[idx]*u.R_sun) > 0)
    color = {1: 'crimson', 0: 'dodgerblue'}[int(color)]
    ax.plot3D(fline.xyz[:,0], fline.xyz[:,1], fline.xyz[:,2], color=color, linewidth=1, alpha=0.4)
ax.set_title('PredSci/MAS Coronal Solution for 2019-04-03')
lim = 3
ax.set_xlim(-lim, lim)
ax.set_ylim(-lim, lim)
ax.set_zlim(-lim, lim)
for axis in 'xyz':
    getattr(ax, 'set_{}label'.format(axis))(axis+' (R$_\odot)$')
plt.show()

Here is the pyvista code in jupyter notebook

mas_output = MASOutput(directory)
br = mas_output['br']

plotter = MASPlotter(mas_output)
nseeds = 12
start_seed = 1.2
r = np.ones(nseeds**2)*start_seed*u.R_sun
lat = np.linspace(-89, 89, nseeds**2, endpoint=False)*u.deg
lon = np.random.rand(nseeds**2)*360*u.deg

# Do the tracing 
max_steps = 10000
step_size = 1
tracer = FortranTracer(max_steps=max_steps, step_size=step_size)
flines = tracer.trace(mas_output, r=r, lat=lat, lon=lon)

for fline in flines:
    # Set color with polarity near the inner boundary 
    color = (br.sample_at_coords(np.mod(fline.lon[0], 2*np.pi*u.rad), fline.lat[0], fline.r[0]*u.R_sun) > 0)
    color = {1: 'crimson', 0: 'dodgerblue'}[int(color)]
    plotter.add_fline(fline, color=color)

# Add a sphere at the inner boundary of the model 
plotter.add_sphere(1*u.R_sun)
plotter.pvplotter.add_axes(x_color='black', y_color='black', z_color='black')
plotter.pvplotter.background_color = 'white'
plotter.show(window_size=[400,400], cpos='yz', jupyter_backend='panel')

Here are the plots from matplotlib and pyvista. Both are done with tracer = FortranTracer(max_steps=10000, step_size=1)

[andyz2021]

Hello,

Thanks for reporting this. In order to help us try and reproduce this, what PsiPy version and carrington rotation were you using for the data?

[AthKouloumvakos]

I have tested this with: './cr2261-high/hmi_mast_mas_std_0101/corona/' and some field lines are weird as shown in the plots above.

My test was with psipy version 0.4.0

import numpy as np
from astropy import units as u
from psipy.model import MASOutput
from psipy.tracing import FortranTracer
from psipy.visualization.pyvista import PyvistaPlotter

mas_output = MASOutput('./cr2261-high/hmi_mast_mas_std_0101/corona/') # location of the predsci/MAS files 
br = mas_output['br']

nseeds = 4
start_seed = 1.2
r = np.ones(nseeds**2)*start_seed*u.R_sun
lat = np.linspace(-89, 89, nseeds**2, endpoint=False)*u.deg
lon = np.random.rand(nseeds**2)*360*u.deg

# Do the tracing 
max_steps = 10000
step_size = 1
tracer = FortranTracer(max_steps=max_steps, step_size=step_size)
flines = tracer.trace(mas_output, r=r, lat=lat, lon=lon)

plotter = PyvistaPlotter(mas_output)
for fline in flines:
    # Set color with polarity near the inner boundary 
    color = (br.sample_at_coords(np.mod(fline.lon[0], 2*np.pi*u.rad), fline.lat[0], fline.r[0]) > 0)
    color = {1: 'crimson', 0: 'dodgerblue'}[int(color)]
    plotter.add_fline(fline, color=color)

# Add a sphere at the inner boundary of the model 
plotter.add_sphere(1*u.R_sun)
plotter.pvplotter.add_axes(x_color='black', y_color='black', z_color='black')
plotter.pvplotter.background_color = 'white'
plotter.show(jupyter_backend='trame') #  jupyter_backend='pythreejs'
# plotter.show(window_size=[400,400], cpos='yz', jupyter_backend='panel')

[andyz2021]

Unfortunately, we're still working on resolving this issue. There's been issues with tracing field lines for both coronal and heliospheric solutions, with the some of the coronal field lines not starting at solar radii 1, as shown below.

Assuming the streamtracer package works as intended, the issue is most likely somewhere in the Python code, maybe in the creation of the VectorGrid or some other data processing step that is somehow messing up the field lines.
I'll keep looking into this and will update if I figure it out.