def run_test(fld, seeds, plot2d=True, plot3d=True, add_title="",
view_kwargs=None, show=False, scatter_mpl=False, mesh_mvi=True):
interpolated_fld = viscid.interp_trilin(fld, seeds)
seed_name = seeds.__class__.__name__
if add_title:
seed_name += " " + add_title
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
# plt.plot(seeds.get_points()[2, :], fld)
mpl_plot_kwargs = dict()
if interpolated_fld.is_spherical():
mpl_plot_kwargs['hemisphere'] = 'north'
vlt.plot(interpolated_fld, **mpl_plot_kwargs)
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
if scatter_mpl:
plt.clf()
vlt.plot2d_line(seeds.get_points(), fld, symdir='z', marker='o')
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
_ = get_mvi_fig(offscreen=not show)
try:
if mesh_mvi:
mesh = vlab.mesh_from_seeds(seeds, scalars=interpolated_fld)
mesh.actor.property.backface_culling = True
except RuntimeError:
pass
pts = seeds.get_points()
p = vlab.points3d(pts[0], pts[1], pts[2], interpolated_fld.flat_data,
scale_mode='none', scale_factor=0.02)
vlab.axes(p)
vlab.title(seed_name)
if view_kwargs:
vlab.view(**view_kwargs)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if show:
vlab.show(stop=True)
except ImportError:
pass
def run_test(_fld, _seeds, plot2d=True, plot3d=True, title='', show=False,
**kwargs):
lines, topo = viscid.calc_streamlines(_fld, _seeds, **kwargs)
topo_color = viscid.topology2color(topo)
# downsample lines for plotting
lines = [line[:, ::8] for line in lines]
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
vlt.plot2d_lines(lines, scalars=topo_color, symdir='y', marker='^')
if title:
plt.title(title)
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
try:
fig = _global_ns['figure']
vlab.clf()
except KeyError:
fig = vlab.figure(size=[1200, 800], offscreen=not show,
bgcolor=(1, 1, 1), fgcolor=(0, 0, 0))
_global_ns['figure'] = fig
fld_mag = np.log(viscid.magnitude(_fld))
try:
# note: mayavi.mlab.mesh can't take color tuples as scalars
# so one can't use topo_color on a mesh surface. This
# is a limitation of mayavi. To actually plot a specific
# set of colors on a mesh, one must use a texture
mesh = vlab.mesh_from_seeds(_seeds, scalars=topo, opacity=0.6)
mesh.actor.property.backface_culling = True
except RuntimeError:
pass
vlab.plot_lines(lines, scalars=fld_mag, tube_radius=0.01,
cmap='viridis')
if title:
vlab.title(title)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if show:
vlab.show()
except ImportError:
pass
def run_test(fld, seeds, plot2d=True, plot3d=True, add_title="",
view_kwargs=None, show=False, scatter_mpl=False, mesh_mvi=True):
interpolated_fld = viscid.interp_trilin(fld, seeds)
seed_name = seeds.__class__.__name__
if add_title:
seed_name += " " + add_title
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
# plt.plot(seeds.get_points()[2, :], fld)
mpl_plot_kwargs = dict()
if interpolated_fld.is_spherical():
mpl_plot_kwargs['hemisphere'] = 'north'
vlt.plot(interpolated_fld, **mpl_plot_kwargs)
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
if scatter_mpl:
plt.clf()
vlt.plot2d_line(seeds.get_points(), fld, symdir='z', marker='o')
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
vlab, _ = get_mvi_fig()
try:
if mesh_mvi:
mesh = vlab.mesh_from_seeds(seeds, scalars=interpolated_fld)
mesh.actor.property.backface_culling = True
except RuntimeError:
pass
pts = seeds.get_points()
p = vlab.points3d(pts[0], pts[1], pts[2], interpolated_fld.flat_data,
scale_mode='none', scale_factor=0.02)
vlab.axes(p)
vlab.title(seed_name)
if view_kwargs:
vlab.view(**view_kwargs)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if show:
vlab.show(stop=True)
except ImportError:
pass
def run_test(fld, seeds, plot2d=True, plot3d=True, add_title="",
view_kwargs=None, show=False):
interpolated_fld = viscid.interp_trilin(fld, seeds)
seed_name = seeds.__class__.__name__
if add_title:
seed_name += " " + add_title
try:
if not plot2d:
raise ImportError
from matplotlib import pyplot as plt
from viscid.plot import vpyplot as vlt
plt.clf()
# plt.plot(seeds.get_points()[2, :], fld)
mpl_plot_kwargs = dict()
if interpolated_fld.is_spherical():
mpl_plot_kwargs['hemisphere'] = 'north'
vlt.plot(interpolated_fld, **mpl_plot_kwargs)
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
try:
fig = _global_ns['figure']
vlab.clf()
except KeyError:
fig = vlab.figure(size=[1200, 800], offscreen=not show)
_global_ns['figure'] = fig
try:
mesh = vlab.mesh_from_seeds(seeds, scalars=interpolated_fld)
mesh.actor.property.backface_culling = True
except RuntimeError:
pass
pts = seeds.get_points()
p = vlab.points3d(pts[0], pts[1], pts[2], interpolated_fld.flat_data,
scale_mode='none', scale_factor=0.02)
vlab.axes(p)
vlab.title(seed_name)
if view_kwargs:
vlab.view(**view_kwargs)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if show:
vlab.show()
except ImportError:
pass
def run_test(_fld, _seeds, plot2d=True, plot3d=True, title='', show=False,
**kwargs):
lines, topo = viscid.calc_streamlines(_fld, _seeds, **kwargs)
topo_color = viscid.topology2color(topo)
# downsample lines for plotting
lines = [line[:, ::8] for line in lines]
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
vlt.plot2d_lines(lines, scalars=topo_color, symdir='y', marker='^')
if title:
plt.title(title)
plt.savefig(next_plot_fname(__file__, series='2d'))
if show:
plt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
vlab, _ = get_mvi_fig()
fld_mag = np.log(viscid.magnitude(_fld))
try:
# note: mayavi.mlab.mesh can't take color tuples as scalars
# so one can't use topo_color on a mesh surface. This
# is a limitation of mayavi. To actually plot a specific
# set of colors on a mesh, one must use a texture
mesh = vlab.mesh_from_seeds(_seeds, scalars=topo, opacity=0.6)
mesh.actor.property.backface_culling = True
except RuntimeError:
pass
vlab.plot_lines(lines, scalars=fld_mag, tube_radius=0.01,
cmap='viridis')
if title:
vlab.title(title)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if show:
vlab.show()
except ImportError:
pass
def _main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--notwo", dest='notwo', action="store_true")
parser.add_argument("--nothree", dest='nothree', action="store_true")
parser.add_argument("--show", "--plot", action="store_true")
args = viscid.vutil.common_argparse(parser, default_verb=0)
plot2d = not args.notwo
plot3d = not args.nothree
# plot2d = True
# plot3d = True
# args.show = True
img = np.load(os.path.join(sample_dir, "logo.npy"))
x = np.linspace(-1, 1, img.shape[0])
y = np.linspace(-1, 1, img.shape[1])
z = np.linspace(-1, 1, img.shape[2])
logo = viscid.arrays2field([x, y, z], img)
if 1:
viscid.logger.info('Testing Point with custom local coordinates...')
pts = np.vstack([[-1, -0.5, 0, 0.5, 1],
[-1, -0.5, 0, 0.5, 1],
[ 0, 0.5, 1, 1.5, 2]])
local_crds = viscid.asarray_datetime64([0, 60, 120, 180, 240],
conservative=True)
seeds = viscid.Point(pts, local_crds=local_crds)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Line...')
seeds = viscid.Line([-1, -1, 0], [1, 1, 2], n=5)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Plane...')
seeds = viscid.Plane([0.0, 0.0, 0.0], [1, 1, 1], [1, 0, 0], 2, 2,
nl=160, nm=170, NL_are_vectors=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Volume...')
seeds = viscid.Volume([-0.8, -0.8, -0.8], [0.8, 0.8, 0.8],
n=[64, 64, 3])
# note: can't make a 2d plot of the volume w/o a slice
run_test(logo, seeds, plot2d=False, plot3d=plot3d, add_title="3d",
show=args.show)
if 1:
viscid.logger.info('Testing Volume (with ignorable dim)...')
seeds = viscid.Volume([-0.8, -0.8, 0.0], [0.8, 0.8, 0.0],
n=[64, 64, 1])
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="2d",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (phi, theta)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (theta, phi)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (phi, theta)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (theta, phi)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Patch...')
seeds = viscid.SphericalPatch(p0=[0, 0, 0], p1=[0, -0, -1],
max_alpha=30.0, max_beta=59.9,
nalpha=65, nbeta=80, r=0.5, roll=45.0)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
# this spline test is very custom
viscid.logger.info('Testing Spline...')
try:
import scipy.interpolate as interpolate
except ImportError:
msg = "XFail: ImportError (is scipy installed?)"
if plot2d:
try:
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
plt.annotate(msg, xy=(0.3, 0.4), xycoords='axes fraction')
plt.savefig(next_plot_fname(__file__, series='2d'))
plt.savefig(next_plot_fname(__file__, series='2d'))
plt.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
plt.show()
except ImportError:
pass
else:
knots = np.array([[ 0.2, 0.5, 0.0], [-0.2, 0.5, 0.2],
[-0.2, 0.0, 0.4], [ 0.2, 0.0, 0.2],
[ 0.2, -0.5, 0.0], [-0.2, -0.5, 0.2]]).T
seed_name = "Spline"
fld = logo
seeds = viscid.Spline(knots)
seed_pts = seeds.get_points()
interp_fld = viscid.interp_trilin(fld, seeds)
if plot2d:
try:
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
vlt.plot(interp_fld)
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
plt.show()
plt.clf()
from matplotlib import rcParams
_ms = rcParams['lines.markersize']
plt.gca().scatter(knots[0, :], knots[1, :],
s=(2 * _ms)**2, marker='^', color='y')
plt.gca().scatter(seed_pts[0, :], seed_pts[1, :],
s=(1.5 * _ms)**2, marker='o', color='k')
vlt.plot2d_line(seed_pts, scalars=interp_fld.flat_data,
symdir='z')
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
plt.show()
except ImportError:
pass
if plot3d:
try:
from viscid.plot import vlab
_ = get_mvi_fig(offscreen=not args.show)
vlab.points3d(knots[0], knots[1], knots[2],
color=(1.0, 1.0, 0), scale_mode='none',
scale_factor=0.04)
p = vlab.points3d(seed_pts[0], seed_pts[1], seed_pts[2],
color=(0, 0, 0), scale_mode='none',
scale_factor=0.03)
vlab.plot_line(seed_pts, scalars=interp_fld.flat_data,
tube_radius=0.01)
vlab.axes(p)
vlab.title(seed_name)
vlab.mlab.roll(-90.0)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
vlab.show(stop=True)
except ImportError:
pass
if 1:
viscid.logger.info('Testing RectilinearMeshPoints...')
f = viscid.load_file(os.path.join(sample_dir, 'sample_xdmf.3d.[-1].xdmf'))
slc = 'x=-40j:12j, y=-10j:10j, z=-10j:10j'
b = f['b'][slc]
z = b.get_crd('z')
sheet_iz = np.argmin(b['x']**2, axis=2)
sheet_pts = b['z=0:1'].get_points()
sheet_pts[2, :] = z[sheet_iz].reshape(-1)
isphere_mask = np.sum(sheet_pts[:2, :]**2, axis=0) < 5**2
day_mask = sheet_pts[0:1, :] > -1.0
sheet_pts[2, :] = np.choose(isphere_mask, [sheet_pts[2, :], 0])
sheet_pts[2, :] = np.choose(day_mask, [sheet_pts[2, :], 0])
nx, ny, _ = b.sshape
sheet_seed = viscid.RectilinearMeshPoints(sheet_pts.reshape(3, nx, ny))
vx_sheet = viscid.interp_nearest(f['vx'], sheet_seed)
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
vlt.clf()
vlt.plot(vx_sheet, symmetric=True)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
vlt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
_ = get_mvi_fig(offscreen=not args.show)
mesh = vlab.mesh_from_seeds(sheet_seed, scalars=vx_sheet,
clim=(-400, 400))
vlab.plot_earth_3d(crd_system=b)
vlab.view(azimuth=+90.0 + 45.0, elevation=90.0 - 25.0,
distance=30.0, focalpoint=(-10.0, +1.0, +1.0))
vlab.title("RectilinearMeshPoints")
vlab.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
vlab.show(stop=True)
except ImportError:
pass
# prevent weird xorg bad-instructions on tear down
if 'figure' in _global_ns and _global_ns['figure'] is not None:
from viscid.plot import vlab
vlab.mlab.close(_global_ns['figure'])
return 0
def _main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--notwo", dest='notwo', action="store_true")
parser.add_argument("--nothree", dest='nothree', action="store_true")
parser.add_argument("--show", "--plot", action="store_true")
args = viscid.vutil.common_argparse(parser, default_verb=0)
plot2d = not args.notwo
plot3d = not args.nothree
# #################################################
# viscid.logger.info("Testing field lines on 2d field...")
B = viscid.make_dipole(twod=True)
line = viscid.seed.Line((0.2, 0.0, 0.0), (1.0, 0.0, 0.0), 10)
obound0 = np.array([-4, -4, -4], dtype=B.data.dtype)
obound1 = np.array([4, 4, 4], dtype=B.data.dtype)
run_test(B,
line,
plot2d=plot2d,
plot3d=plot3d,
title='2D',
show=args.show,
ibound=0.07,
obound0=obound0,
obound1=obound1)
#################################################
viscid.logger.info("Testing field lines on 3d field...")
B = viscid.make_dipole(m=[0.2, 0.3, -0.9])
sphere = viscid.seed.Sphere((0.0, 0.0, 0.0), 2.0, ntheta=20, nphi=10)
obound0 = np.array([-4, -4, -4], dtype=B.data.dtype)
obound1 = np.array([4, 4, 4], dtype=B.data.dtype)
run_test(B,
sphere,
plot2d=plot2d,
plot3d=plot3d,
title='3D',
show=args.show,
ibound=0.12,
obound0=obound0,
obound1=obound1,
method=viscid.RK12)
# The Remainder of this test makes sure higher order methods are indeed
# more accurate than lower order methods... this could find a bug in
# the integrators
##################################################
# test accuracy of streamlines in an ideal dipole
cotr = viscid.Cotr(dip_tilt=15.0, dip_gsm=21.0) # pylint: disable=not-callable
m = cotr.get_dipole_moment(crd_system='gse')
seeds = viscid.seed.Sphere((0.0, 0.0, 0.0),
2.0,
pole=-m,
ntheta=25,
nphi=25,
thetalim=(5, 90),
philim=(5, 360),
phi_endpoint=False)
B = viscid.make_dipole(m=m,
crd_system='gse',
n=(256, 256, 256),
l=(-25, -25, -25),
h=(25, 25, 25),
dtype='f8')
seeds_xyz = seeds.get_points()
# seeds_lsp = viscid.xyz2lsrlp(seeds_xyz, cotr=cotr, crd_system=B)[(0, 3), :]
seeds_lsp = viscid.xyz2lsrlp(seeds_xyz, cotr=cotr, crd_system=B)[(0, 3), :]
e1_lines, e1_lsps, t_e1 = lines_and_lsps(B,
seeds,
method='euler1',
ibound=1.0,
cotr=cotr)
rk2_lines, rk2_lsps, t_rk2 = lines_and_lsps(B,
seeds,
method='rk2',
ibound=1.0,
cotr=cotr)
rk4_lines, rk4_lsps, t_rk4 = lines_and_lsps(B,
seeds,
method='rk4',
ibound=1.0,
cotr=cotr)
e1a_lines, e1a_lsps, t_e1a = lines_and_lsps(B,
seeds,
method='euler1a',
ibound=1.0,
cotr=cotr)
rk12_lines, rk12_lsps, t_rk12 = lines_and_lsps(B,
seeds,
method='rk12',
ibound=1.0,
cotr=cotr)
rk45_lines, rk45_lsps, t_rk45 = lines_and_lsps(B,
seeds,
method='rk45',
ibound=1.0,
cotr=cotr)
def _calc_rel_diff(_lsp, _ideal_lsp, _d):
_diffs = []
for _ilsp, _iideal in zip(_lsp, _ideal_lsp.T):
_a = _ilsp[_d, :]
_b = _iideal[_d]
_diffs.append((_a - _b) / _b)
return _diffs
lshell_diff_e1 = _calc_rel_diff(e1_lsps, seeds_lsp, 0)
phi_diff_e1 = _calc_rel_diff(e1_lsps, seeds_lsp, 1)
lshell_diff_rk2 = _calc_rel_diff(rk2_lsps, seeds_lsp, 0)
phi_diff_rk2 = _calc_rel_diff(rk2_lsps, seeds_lsp, 1)
lshell_diff_rk4 = _calc_rel_diff(rk4_lsps, seeds_lsp, 0)
phi_diff_rk4 = _calc_rel_diff(rk4_lsps, seeds_lsp, 1)
lshell_diff_e1a = _calc_rel_diff(e1a_lsps, seeds_lsp, 0)
phi_diff_e1a = _calc_rel_diff(e1a_lsps, seeds_lsp, 1)
lshell_diff_rk12 = _calc_rel_diff(rk12_lsps, seeds_lsp, 0)
phi_diff_rk12 = _calc_rel_diff(rk12_lsps, seeds_lsp, 1)
lshell_diff_rk45 = _calc_rel_diff(rk45_lsps, seeds_lsp, 0)
phi_diff_rk45 = _calc_rel_diff(rk45_lsps, seeds_lsp, 1)
methods = [
'Euler 1', 'Runge Kutta 2', 'Runge Kutta 4', 'Euler 1 Adaptive Step',
'Runge Kutta 12 Adaptive Step', 'Runge Kutta 45 Adaptive Step'
]
wall_ts = [t_e1, t_rk2, t_rk4, t_e1a, t_rk12, t_rk45]
all_lines = [
e1_lines, rk2_lines, rk4_lines, e1a_lines, rk12_lines, rk45_lines
]
all_lshell_diffs = [
lshell_diff_e1, lshell_diff_rk2, lshell_diff_rk4, lshell_diff_e1a,
lshell_diff_rk12, lshell_diff_rk45
]
lshell_diffs = [
np.abs(np.concatenate(lshell_diff_e1, axis=0)),
np.abs(np.concatenate(lshell_diff_rk2, axis=0)),
np.abs(np.concatenate(lshell_diff_rk4, axis=0)),
np.abs(np.concatenate(lshell_diff_e1a, axis=0)),
np.abs(np.concatenate(lshell_diff_rk12, axis=0)),
np.abs(np.concatenate(lshell_diff_rk45, axis=0))
]
phi_diffs = [
np.abs(np.concatenate(phi_diff_e1, axis=0)),
np.abs(np.concatenate(phi_diff_rk2, axis=0)),
np.abs(np.concatenate(phi_diff_rk4, axis=0)),
np.abs(np.concatenate(phi_diff_e1a, axis=0)),
np.abs(np.concatenate(phi_diff_rk12, axis=0)),
np.abs(np.concatenate(phi_diff_rk45, axis=0))
]
npts = [len(lsd) for lsd in lshell_diffs]
lshell_75 = [np.percentile(lsdiff, 75) for lsdiff in lshell_diffs]
# # 3D DEBUG PLOT:: for really getting under the covers
# vlab.clf()
# earth1 = viscid.seed.Sphere((0.0, 0.0, 0.0), 1.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls1 = viscid.xyz2lsrlp(earth1.get_points(), cotr=cotr, crd_system='gse')[0, :]
# earth2 = viscid.seed.Sphere((0.0, 0.0, 0.0), 2.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls2 = viscid.xyz2lsrlp(earth2.get_points(), cotr=cotr, crd_system='gse')[0, :]
# earth4 = viscid.seed.Sphere((0.0, 0.0, 0.0), 4.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls4 = viscid.xyz2lsrlp(earth4.get_points(), cotr=cotr, crd_system='gse')[0, :]
# clim = [2.0, 6.0]
# vlab.mesh_from_seeds(earth1, scalars=ls1, clim=clim, logscale=True)
# vlab.mesh_from_seeds(earth2, scalars=ls2, clim=clim, logscale=True, opacity=0.5)
# vlab.mesh_from_seeds(earth4, scalars=ls2, clim=clim, logscale=True, opacity=0.25)
# vlab.plot3d_lines(e1_lines, scalars=[_e1_lsp[0, :] for _e1_lsp in e1_lsps],
# clim=clim, logscale=True)
# vlab.colorbar(title="L-Shell")
# vlab.show()
assert lshell_75[1] < lshell_75[0], "RK2 should have less error than Euler"
assert lshell_75[2] < lshell_75[1], "RK4 should have less error than RK2"
assert lshell_75[3] < lshell_75[
0], "Euler 1a should have less error than Euler 1"
assert lshell_75[4] < lshell_75[
0], "RK 12 should have less error than Euler 1"
assert lshell_75[5] < lshell_75[1], "RK 45 should have less error than RK2"
try:
if not plot2d:
raise ImportError
from matplotlib import pyplot as plt
from viscid.plot import vpyplot as vlt
# stats on error for all points on all lines
_ = plt.figure(figsize=(15, 8))
ax1 = vlt.subplot(121)
v = plt.violinplot(lshell_diffs,
showextrema=False,
showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
t_txt = ", took {0:.2e} seconds".format(wall_ts[i])
stat_txt = format_data_range(lshell_diffs[i])
plt.text(xl + 0.35 * (xh - xl), i + 1.15, txt + t_txt, color=c)
plt.text(xl + 0.35 * (xh - xl), i + 0.85, stat_txt, color=c)
ax1.get_yaxis().set_visible(False)
plt.title('L-Shell')
plt.xlabel('Relative Difference from Ideal (as fraction)')
ax2 = vlt.subplot(122)
v = plt.violinplot(phi_diffs,
showextrema=False,
showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
t_txt = ", took {0:.2e} seconds".format(wall_ts[i])
stat_txt = format_data_range(phi_diffs[i])
plt.text(xl + 0.35 * (xh - xl), i + 1.15, txt + t_txt, color=c)
plt.text(xl + 0.35 * (xh - xl), i + 0.85, stat_txt, color=c)
ax2.get_yaxis().set_visible(False)
plt.title('Longitude')
plt.xlabel('Relative Difference from Ideal (as fraction)')
vlt.auto_adjust_subplots()
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
# stats for ds for all points on all lines
_ = plt.figure(figsize=(10, 8))
ax1 = vlt.subplot(111)
ds = [
np.concatenate([
np.linalg.norm(_l[:, 1:] - _l[:, :-1], axis=0) for _l in lines
]) for lines in all_lines
]
v = plt.violinplot(ds,
showextrema=False,
showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
stat_txt = format_data_range(ds[i])
plt.text(xl + 0.01 * (xh - xl), i + 1.15, txt, color=c)
plt.text(xl + 0.01 * (xh - xl), i + 0.85, stat_txt, color=c)
ax1.get_yaxis().set_visible(False)
plt.xscale('log')
plt.title('Step Size')
plt.xlabel('Absolute Step Size')
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
# random other information
_ = plt.figure(figsize=(13, 10))
## wall time for each method
vlt.subplot(221)
plt.scatter(range(len(methods)),
wall_ts,
color=colors,
s=150,
marker='s',
edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, wall_ts[i]),
xytext=(0, 15.0),
color=colors[i],
horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Wall Time (s)")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(wall_ts), np.max(wall_ts)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## number of points calculated for each method
vlt.subplot(222)
plt.scatter(range(len(methods)),
npts,
color=colors,
s=150,
marker='s',
edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, npts[i]),
xytext=(0, 15.0),
color=colors[i],
horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Number of Streamline Points Calculated")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(npts), np.max(npts)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## Wall time per segment, this should show the overhead of the method
vlt.subplot(223)
wall_t_per_seg = np.asarray(wall_ts) / np.asarray(npts)
plt.scatter(range(len(methods)),
wall_t_per_seg,
color=colors,
s=150,
marker='s',
edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, wall_t_per_seg[i]),
xytext=(0, 15.0),
color=colors[i],
horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Wall Time Per Line Segment")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(wall_t_per_seg), np.max(wall_t_per_seg)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
plt.gca().xaxis.set_major_formatter(viscid.plot.mpl_extra.steve_axfmt)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## 75th percentile of l-shell error for each method
vlt.subplot(224)
plt.scatter(range(len(methods)),
lshell_75,
color=colors,
s=150,
marker='s',
edgecolors='none')
plt.yscale('log')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, lshell_75[i]),
xytext=(0, 15.0),
color=colors[i],
horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("75th Percentile of Relative L-Shell Error")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
ymin, ymax = np.min(lshell_75), np.max(lshell_75)
plt.ylim(0.75 * ymin, 2.5 * ymax)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
vlt.auto_adjust_subplots(subplot_params=dict(wspace=0.25, hspace=0.15))
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
try:
fig = _global_ns['figure']
vlab.clf()
except KeyError:
fig = vlab.figure(size=[1200, 800],
offscreen=not args.show,
bgcolor=(1, 1, 1),
fgcolor=(0, 0, 0))
_global_ns['figure'] = fig
for i, method in zip(count(), methods):
# if i in (3, 4):
# next_plot_fname(__file__, series='q3')
# print(i, "::", [line.shape[1] for line in all_lines[i]])
# # continue
vlab.clf()
_lshell_diff = [np.abs(s) for s in all_lshell_diffs[i]]
vlab.plot3d_lines(all_lines[i], scalars=_lshell_diff)
vlab.colorbar(title="Relative L-Shell Error (as fraction)")
vlab.title(method, size=0.5)
vlab.orientation_axes()
vlab.view(azimuth=40,
elevation=140,
distance=80.0,
focalpoint=[0, 0, 0])
vlab.savefig(next_plot_fname(__file__, series='q3'))
if args.show:
vlab.show()
except ImportError:
pass
# prevent weird xorg bad-instructions on tear down
if 'figure' in _global_ns and _global_ns['figure'] is not None:
from viscid.plot import vlab
vlab.mlab.close(_global_ns['figure'])
return 0
def _main():
global offscreen_vlab
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--notwo", dest='notwo', action="store_true")
parser.add_argument("--nothree", dest='nothree', action="store_true")
parser.add_argument("--show", "--plot", action="store_true")
args = viscid.vutil.common_argparse(parser, default_verb=0)
plot2d = not args.notwo
plot3d = not args.nothree
# plot2d = True
# plot3d = True
# args.show = True
offscreen_vlab = not args.show
img = np.load(os.path.join(sample_dir, "logo.npy"))
x = np.linspace(-1, 1, img.shape[0])
y = np.linspace(-1, 1, img.shape[1])
z = np.linspace(-1, 1, img.shape[2])
logo = viscid.arrays2field([x, y, z], img)
if 1:
viscid.logger.info('Testing Point with custom local coordinates...')
pts = np.vstack([[-1, -0.5, 0, 0.5, 1],
[-1, -0.5, 0, 0.5, 1],
[ 0, 0.5, 1, 1.5, 2]])
local_crds = viscid.asarray_datetime64([0, 60, 120, 180, 240],
conservative=True)
seeds = viscid.Point(pts, local_crds=local_crds)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Line...')
seeds = viscid.Line([-1, -1, 0], [1, 1, 2], n=5)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Plane...')
seeds = viscid.Plane([0.0, 0.0, 0.0], [1, 1, 1], [1, 0, 0], 2, 2,
nl=160, nm=170, NL_are_vectors=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Volume...')
seeds = viscid.Volume([-0.8, -0.8, -0.8], [0.8, 0.8, 0.8],
n=[64, 64, 3])
# note: can't make a 2d plot of the volume w/o a slice
run_test(logo, seeds, plot2d=False, plot3d=plot3d, add_title="3d",
show=args.show)
if 1:
viscid.logger.info('Testing Volume (with ignorable dim)...')
seeds = viscid.Volume([-0.8, -0.8, 0.0], [0.8, 0.8, 0.0],
n=[64, 64, 1])
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="2d",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (phi, theta)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (theta, phi)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (phi, theta)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (theta, phi)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Patch...')
seeds = viscid.SphericalPatch(p0=[0, 0, 0], p1=[0, -0, -1],
max_alpha=30.0, max_beta=59.9,
nalpha=65, nbeta=80, r=0.5, roll=45.0)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
# this spline test is very custom
viscid.logger.info('Testing Spline...')
try:
import scipy.interpolate as interpolate
except ImportError:
msg = "XFail: ImportError (is scipy installed?)"
if plot2d:
try:
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
plt.annotate(msg, xy=(0.3, 0.4), xycoords='axes fraction')
plt.savefig(next_plot_fname(__file__, series='2d'))
plt.savefig(next_plot_fname(__file__, series='2d'))
plt.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
plt.show()
except ImportError:
pass
else:
knots = np.array([[ 0.2, 0.5, 0.0], [-0.2, 0.5, 0.2],
[-0.2, 0.0, 0.4], [ 0.2, 0.0, 0.2],
[ 0.2, -0.5, 0.0], [-0.2, -0.5, 0.2]]).T
seed_name = "Spline"
fld = logo
seeds = viscid.Spline(knots)
seed_pts = seeds.get_points()
interp_fld = viscid.interp_trilin(fld, seeds)
if plot2d:
try:
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
plt.clf()
vlt.plot(interp_fld)
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
plt.show()
plt.clf()
from matplotlib import rcParams
_ms = rcParams['lines.markersize']
plt.gca().scatter(knots[0, :], knots[1, :],
s=(2 * _ms)**2, marker='^', color='y')
plt.gca().scatter(seed_pts[0, :], seed_pts[1, :],
s=(1.5 * _ms)**2, marker='o', color='k')
vlt.plot2d_line(seed_pts, scalars=interp_fld.flat_data,
symdir='z')
plt.title(seed_name)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
plt.show()
except ImportError:
pass
if plot3d:
try:
vlab, _ = get_mvi_fig()
vlab.points3d(knots[0], knots[1], knots[2],
color=(1.0, 1.0, 0), scale_mode='none',
scale_factor=0.04)
p = vlab.points3d(seed_pts[0], seed_pts[1], seed_pts[2],
color=(0, 0, 0), scale_mode='none',
scale_factor=0.03)
vlab.plot_line(seed_pts, scalars=interp_fld.flat_data,
tube_radius=0.01)
vlab.axes(p)
vlab.title(seed_name)
vlab.mlab.roll(-90.0)
vlab.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
vlab.show(stop=True)
except ImportError:
pass
if 1:
viscid.logger.info('Testing RectilinearMeshPoints...')
f = viscid.load_file(os.path.join(sample_dir, 'sample_xdmf.3d.[-1].xdmf'))
slc = 'x=-40j:12j, y=-10j:10j, z=-10j:10j'
b = f['b'][slc]
z = b.get_crd('z')
sheet_iz = np.argmin(b['x']**2, axis=2)
sheet_pts = b['z=0:1'].get_points()
sheet_pts[2, :] = z[sheet_iz].reshape(-1)
isphere_mask = np.sum(sheet_pts[:2, :]**2, axis=0) < 5**2
day_mask = sheet_pts[0:1, :] > -1.0
sheet_pts[2, :] = np.choose(isphere_mask, [sheet_pts[2, :], 0])
sheet_pts[2, :] = np.choose(day_mask, [sheet_pts[2, :], 0])
nx, ny, _ = b.sshape
sheet_seed = viscid.RectilinearMeshPoints(sheet_pts.reshape(3, nx, ny))
vx_sheet = viscid.interp_nearest(f['vx'], sheet_seed)
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
vlt.clf()
vlt.plot(vx_sheet, symmetric=True)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
vlt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
vlab, _ = get_mvi_fig()
mesh = vlab.mesh_from_seeds(sheet_seed, scalars=vx_sheet,
clim=(-400, 400))
vlab.plot_earth_3d(crd_system=b)
vlab.view(azimuth=+90.0 + 45.0, elevation=90.0 - 25.0,
distance=30.0, focalpoint=(-10.0, +1.0, +1.0))
vlab.title("RectilinearMeshPoints")
vlab.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
vlab.show(stop=True)
except ImportError:
pass
# prevent weird xorg bad-instructions on tear down
if 'figure' in _global_ns and _global_ns['figure'] is not None:
from viscid.plot import vlab
vlab.mlab.close(_global_ns['figure'])
return 0
def _main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--notwo", dest='notwo', action="store_true")
parser.add_argument("--nothree", dest='nothree', action="store_true")
parser.add_argument("--show", "--plot", action="store_true")
args = viscid.vutil.common_argparse(parser, default_verb=0)
plot2d = not args.notwo
plot3d = not args.nothree
# plot2d = True
# plot3d = True
# args.show = True
img = np.load(os.path.join(sample_dir, "logo.npy"))
x = np.linspace(-1, 1, img.shape[0])
y = np.linspace(-1, 1, img.shape[1])
z = np.linspace(-1, 1, img.shape[2])
logo = viscid.arrays2field([x, y, z], img)
if 1:
viscid.logger.info('Testing Line...')
seeds = viscid.Line([-1, -1, 0], [1, 1, 2], n=5)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Plane...')
seeds = viscid.Plane([0.0, 0.0, 0.0], [1, 1, 1], [1, 0, 0], 2, 2,
nl=160, nm=170, NL_are_vectors=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing Volume...')
seeds = viscid.Volume([-0.8, -0.8, -0.8], [0.8, 0.8, 0.8],
n=[64, 64, 3])
# note: can't make a 2d plot of the volume w/o a slice
run_test(logo, seeds, plot2d=False, plot3d=plot3d, add_title="3d",
show=args.show)
if 1:
viscid.logger.info('Testing Volume (with ignorable dim)...')
seeds = viscid.Volume([-0.8, -0.8, 0.0], [0.8, 0.8, 0.0],
n=[64, 64, 1])
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="2d",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (phi, theta)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Sphere (theta, phi)...')
seeds = viscid.Sphere([0, 0, 0], r=1.0, ntheta=160, nphi=170,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (phi, theta)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=False)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="PT",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Cap (theta, phi)...')
seeds = viscid.SphericalCap(p0=[0, 0, 0], r=1.0, ntheta=64, nphi=80,
pole=[-1, -1, -1], theta_phi=True)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, add_title="TP",
view_kwargs=dict(azimuth=180, elevation=180), show=args.show)
if 1:
viscid.logger.info('Testing Spherical Patch...')
seeds = viscid.SphericalPatch(p0=[0, 0, 0], p1=[0, -0, -1],
max_alpha=30.0, max_beta=59.9,
nalpha=65, nbeta=80, r=0.5, roll=45.0)
run_test(logo, seeds, plot2d=plot2d, plot3d=plot3d, show=args.show)
if 1:
viscid.logger.info('Testing RectilinearMeshPoints...')
f = viscid.load_file(os.path.join(sample_dir, 'sample_xdmf.3d.[-1].xdmf'))
slc = 'x=-40f:12f, y=-10f:10f, z=-10f:10f'
b = f['b'][slc]
z = b.get_crd('z')
sheet_iz = np.argmin(b['x']**2, axis=2)
sheet_pts = b['z=0:1'].get_points()
sheet_pts[2, :] = z[sheet_iz].reshape(-1)
isphere_mask = np.sum(sheet_pts[:2, :]**2, axis=0) < 5**2
day_mask = sheet_pts[0:1, :] > -1.0
sheet_pts[2, :] = np.choose(isphere_mask, [sheet_pts[2, :], 0])
sheet_pts[2, :] = np.choose(day_mask, [sheet_pts[2, :], 0])
nx, ny, _ = b.sshape
sheet_seed = viscid.RectilinearMeshPoints(sheet_pts.reshape(3, nx, ny))
vx_sheet = viscid.interp_nearest(f['vx'], sheet_seed)
try:
if not plot2d:
raise ImportError
from matplotlib import pyplot as plt
from viscid.plot import vpyplot as vlt
vlt.clf()
vlt.plot(vx_sheet, symmetric=True)
plt.savefig(next_plot_fname(__file__, series='2d'))
if args.show:
vlt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
vlab.clf()
mesh = vlab.mesh_from_seeds(sheet_seed, scalars=vx_sheet,
clim=(-400, 400))
vlab.plot_earth_3d(crd_system=b)
vlab.view(azimuth=+90.0 + 45.0, elevation=90.0 - 25.0,
distance=30.0, focalpoint=(-10.0, +1.0, +1.0))
vlab.title("RectilinearMeshPoints")
vlab.savefig(next_plot_fname(__file__, series='3d'))
if args.show:
vlab.show()
except ImportError:
pass
# prevent weird xorg bad-instructions on tear down
if 'figure' in _global_ns and _global_ns['figure'] is not None:
from viscid.plot import vlab
vlab.mlab.close(_global_ns['figure'])
return 0
def _main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--notwo", dest='notwo', action="store_true")
parser.add_argument("--nothree", dest='nothree', action="store_true")
parser.add_argument("--show", "--plot", action="store_true")
args = viscid.vutil.common_argparse(parser, default_verb=0)
plot2d = not args.notwo
plot3d = not args.nothree
# #################################################
# viscid.logger.info("Testing field lines on 2d field...")
B = viscid.make_dipole(twod=True)
line = viscid.seed.Line((0.2, 0.0, 0.0), (1.0, 0.0, 0.0), 10)
obound0 = np.array([-4, -4, -4], dtype=B.data.dtype)
obound1 = np.array([4, 4, 4], dtype=B.data.dtype)
run_test(B, line, plot2d=plot2d, plot3d=plot3d, title='2D', show=args.show,
ibound=0.07, obound0=obound0, obound1=obound1)
#################################################
viscid.logger.info("Testing field lines on 3d field...")
B = viscid.make_dipole(m=[0.2, 0.3, -0.9])
sphere = viscid.seed.Sphere((0.0, 0.0, 0.0), 2.0, ntheta=20, nphi=10)
obound0 = np.array([-4, -4, -4], dtype=B.data.dtype)
obound1 = np.array([4, 4, 4], dtype=B.data.dtype)
run_test(B, sphere, plot2d=plot2d, plot3d=plot3d, title='3D', show=args.show,
ibound=0.12, obound0=obound0, obound1=obound1, method=viscid.RK12)
# The Remainder of this test makes sure higher order methods are indeed
# more accurate than lower order methods... this could find a bug in
# the integrators
##################################################
# test accuracy of streamlines in an ideal dipole
cotr = viscid.Cotr(dip_tilt=15.0, dip_gsm=21.0) # pylint: disable=not-callable
m = cotr.get_dipole_moment(crd_system='gse')
seeds = viscid.seed.Sphere((0.0, 0.0, 0.0), 2.0, pole=-m, ntheta=25, nphi=25,
thetalim=(5, 90), philim=(5, 360), phi_endpoint=False)
B = viscid.make_dipole(m=m, crd_system='gse', n=(256, 256, 256),
l=(-25, -25, -25), h=(25, 25, 25), dtype='f8')
seeds_xyz = seeds.get_points()
# seeds_lsp = viscid.xyz2lsrlp(seeds_xyz, cotr=cotr, crd_system=B)[(0, 3), :]
seeds_lsp = viscid.xyz2lsrlp(seeds_xyz, cotr=cotr, crd_system=B)[(0, 3), :]
e1_lines, e1_lsps, t_e1 = lines_and_lsps(B, seeds, method='euler1',
ibound=1.0, cotr=cotr)
rk2_lines, rk2_lsps, t_rk2 = lines_and_lsps(B, seeds, method='rk2',
ibound=1.0, cotr=cotr)
rk4_lines, rk4_lsps, t_rk4 = lines_and_lsps(B, seeds, method='rk4',
ibound=1.0, cotr=cotr)
e1a_lines, e1a_lsps, t_e1a = lines_and_lsps(B, seeds, method='euler1a',
ibound=1.0, cotr=cotr)
rk12_lines, rk12_lsps, t_rk12 = lines_and_lsps(B, seeds, method='rk12',
ibound=1.0, cotr=cotr)
rk45_lines, rk45_lsps, t_rk45 = lines_and_lsps(B, seeds, method='rk45',
ibound=1.0, cotr=cotr)
def _calc_rel_diff(_lsp, _ideal_lsp, _d):
_diffs = []
for _ilsp, _iideal in zip(_lsp, _ideal_lsp.T):
_a = _ilsp[_d, :]
_b = _iideal[_d]
_diffs.append((_a - _b) / _b)
return _diffs
lshell_diff_e1 = _calc_rel_diff(e1_lsps, seeds_lsp, 0)
phi_diff_e1 = _calc_rel_diff(e1_lsps, seeds_lsp, 1)
lshell_diff_rk2 = _calc_rel_diff(rk2_lsps, seeds_lsp, 0)
phi_diff_rk2 = _calc_rel_diff(rk2_lsps, seeds_lsp, 1)
lshell_diff_rk4 = _calc_rel_diff(rk4_lsps, seeds_lsp, 0)
phi_diff_rk4 = _calc_rel_diff(rk4_lsps, seeds_lsp, 1)
lshell_diff_e1a = _calc_rel_diff(e1a_lsps, seeds_lsp, 0)
phi_diff_e1a = _calc_rel_diff(e1a_lsps, seeds_lsp, 1)
lshell_diff_rk12 = _calc_rel_diff(rk12_lsps, seeds_lsp, 0)
phi_diff_rk12 = _calc_rel_diff(rk12_lsps, seeds_lsp, 1)
lshell_diff_rk45 = _calc_rel_diff(rk45_lsps, seeds_lsp, 0)
phi_diff_rk45 = _calc_rel_diff(rk45_lsps, seeds_lsp, 1)
methods = ['Euler 1', 'Runge Kutta 2', 'Runge Kutta 4',
'Euler 1 Adaptive Step', 'Runge Kutta 12 Adaptive Step',
'Runge Kutta 45 Adaptive Step']
wall_ts = [t_e1, t_rk2, t_rk4, t_e1a, t_rk12, t_rk45]
all_lines = [e1_lines, rk2_lines, rk4_lines, e1a_lines, rk12_lines,
rk45_lines]
all_lshell_diffs = [lshell_diff_e1, lshell_diff_rk2, lshell_diff_rk4,
lshell_diff_e1a, lshell_diff_rk12, lshell_diff_rk45]
lshell_diffs = [np.abs(np.concatenate(lshell_diff_e1, axis=0)),
np.abs(np.concatenate(lshell_diff_rk2, axis=0)),
np.abs(np.concatenate(lshell_diff_rk4, axis=0)),
np.abs(np.concatenate(lshell_diff_e1a, axis=0)),
np.abs(np.concatenate(lshell_diff_rk12, axis=0)),
np.abs(np.concatenate(lshell_diff_rk45, axis=0))]
phi_diffs = [np.abs(np.concatenate(phi_diff_e1, axis=0)),
np.abs(np.concatenate(phi_diff_rk2, axis=0)),
np.abs(np.concatenate(phi_diff_rk4, axis=0)),
np.abs(np.concatenate(phi_diff_e1a, axis=0)),
np.abs(np.concatenate(phi_diff_rk12, axis=0)),
np.abs(np.concatenate(phi_diff_rk45, axis=0))]
npts = [len(lsd) for lsd in lshell_diffs]
lshell_75 = [np.percentile(lsdiff, 75) for lsdiff in lshell_diffs]
# # 3D DEBUG PLOT:: for really getting under the covers
# vlab.clf()
# earth1 = viscid.seed.Sphere((0.0, 0.0, 0.0), 1.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls1 = viscid.xyz2lsrlp(earth1.get_points(), cotr=cotr, crd_system='gse')[0, :]
# earth2 = viscid.seed.Sphere((0.0, 0.0, 0.0), 2.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls2 = viscid.xyz2lsrlp(earth2.get_points(), cotr=cotr, crd_system='gse')[0, :]
# earth4 = viscid.seed.Sphere((0.0, 0.0, 0.0), 4.0, pole=-m, ntheta=60, nphi=120,
# thetalim=(15, 165), philim=(0, 360))
# ls4 = viscid.xyz2lsrlp(earth4.get_points(), cotr=cotr, crd_system='gse')[0, :]
# clim = [2.0, 6.0]
# vlab.mesh_from_seeds(earth1, scalars=ls1, clim=clim, logscale=True)
# vlab.mesh_from_seeds(earth2, scalars=ls2, clim=clim, logscale=True, opacity=0.5)
# vlab.mesh_from_seeds(earth4, scalars=ls2, clim=clim, logscale=True, opacity=0.25)
# vlab.plot3d_lines(e1_lines, scalars=[_e1_lsp[0, :] for _e1_lsp in e1_lsps],
# clim=clim, logscale=True)
# vlab.colorbar(title="L-Shell")
# vlab.show()
assert lshell_75[1] < lshell_75[0], "RK2 should have less error than Euler"
assert lshell_75[2] < lshell_75[1], "RK4 should have less error than RK2"
assert lshell_75[3] < lshell_75[0], "Euler 1a should have less error than Euler 1"
assert lshell_75[4] < lshell_75[0], "RK 12 should have less error than Euler 1"
assert lshell_75[5] < lshell_75[1], "RK 45 should have less error than RK2"
try:
if not plot2d:
raise ImportError
from viscid.plot import vpyplot as vlt
from matplotlib import pyplot as plt
# stats on error for all points on all lines
_ = plt.figure(figsize=(15, 8))
ax1 = vlt.subplot(121)
v = plt.violinplot(lshell_diffs, showextrema=False, showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
t_txt = ", took {0:.2e} seconds".format(wall_ts[i])
stat_txt = format_data_range(lshell_diffs[i])
plt.text(xl + 0.35 * (xh - xl), i + 1.15, txt + t_txt, color=c)
plt.text(xl + 0.35 * (xh - xl), i + 0.85, stat_txt, color=c)
ax1.get_yaxis().set_visible(False)
plt.title('L-Shell')
plt.xlabel('Relative Difference from Ideal (as fraction)')
ax2 = vlt.subplot(122)
v = plt.violinplot(phi_diffs, showextrema=False, showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
t_txt = ", took {0:.2e} seconds".format(wall_ts[i])
stat_txt = format_data_range(phi_diffs[i])
plt.text(xl + 0.35 * (xh - xl), i + 1.15, txt + t_txt, color=c)
plt.text(xl + 0.35 * (xh - xl), i + 0.85, stat_txt, color=c)
ax2.get_yaxis().set_visible(False)
plt.title('Longitude')
plt.xlabel('Relative Difference from Ideal (as fraction)')
vlt.auto_adjust_subplots()
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
# stats for ds for all points on all lines
_ = plt.figure(figsize=(10, 8))
ax1 = vlt.subplot(111)
ds = [np.concatenate([np.linalg.norm(_l[:, 1:] - _l[:, :-1], axis=0)
for _l in lines]) for lines in all_lines]
v = plt.violinplot(ds, showextrema=False, showmedians=False,
vert=False)
colors = set_violin_colors(v)
xl, xh = plt.gca().get_xlim()
for i, txt, c in zip(count(), methods, colors):
stat_txt = format_data_range(ds[i])
plt.annotate(txt, xy=(0.55, i / len(methods) + 0.1), color=c,
xycoords='axes fraction')
plt.annotate(stat_txt, xy=(0.55, i / len(methods) + 0.04), color=c,
xycoords='axes fraction')
ax1.get_yaxis().set_visible(False)
plt.xscale('log')
plt.title('Step Size')
plt.xlabel('Absolute Step Size')
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
# random other information
_ = plt.figure(figsize=(13, 10))
## wall time for each method
vlt.subplot(221)
plt.scatter(range(len(methods)), wall_ts, color=colors,
s=150, marker='s', edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, wall_ts[i]), xytext=(0, 15.0),
color=colors[i], horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Wall Time (s)")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(wall_ts), np.max(wall_ts)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## number of points calculated for each method
vlt.subplot(222)
plt.scatter(range(len(methods)), npts, color=colors,
s=150, marker='s', edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, npts[i]), xytext=(0, 15.0),
color=colors[i], horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Number of Streamline Points Calculated")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(npts), np.max(npts)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## Wall time per segment, this should show the overhead of the method
vlt.subplot(223)
wall_t_per_seg = np.asarray(wall_ts) / np.asarray(npts)
plt.scatter(range(len(methods)), wall_t_per_seg, color=colors,
s=150, marker='s', edgecolors='none')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, wall_t_per_seg[i]), xytext=(0, 15.0),
color=colors[i], horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("Wall Time Per Line Segment")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
yl, yh = np.min(wall_t_per_seg), np.max(wall_t_per_seg)
y_padding = 0.4 * (yh - yl)
plt.ylim(yl - y_padding, yh + y_padding)
plt.gca().get_xaxis().set_visible(False)
plt.gca().xaxis.set_major_formatter(viscid.plot.mpl_extra.steve_axfmt)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
## 75th percentile of l-shell error for each method
vlt.subplot(224)
plt.scatter(range(len(methods)), lshell_75, color=colors,
s=150, marker='s', edgecolors='none')
plt.yscale('log')
for i, meth in enumerate(methods):
meth = meth.replace(" Adaptive Step", "\nAdaptive Step")
plt.annotate(meth, (i, lshell_75[i]), xytext=(0, 15.0),
color=colors[i], horizontalalignment='center',
verticalalignment='bottom',
textcoords='offset points')
plt.ylabel("75th Percentile of Relative L-Shell Error")
x_padding = 0.5
plt.xlim(-x_padding, len(methods) - x_padding)
ymin, ymax = np.min(lshell_75), np.max(lshell_75)
plt.ylim(0.75 * ymin, 2.5 * ymax)
plt.gca().get_xaxis().set_visible(False)
for _which in ('right', 'top'):
plt.gca().spines[_which].set_color('none')
vlt.auto_adjust_subplots(subplot_params=dict(wspace=0.25, hspace=0.15))
vlt.savefig(next_plot_fname(__file__, series='q2'))
if args.show:
vlt.show()
except ImportError:
pass
try:
if not plot3d:
raise ImportError
from viscid.plot import vlab
try:
fig = _global_ns['figure']
vlab.clf()
except KeyError:
fig = vlab.figure(size=[1200, 800], offscreen=not args.show,
bgcolor=(1, 1, 1), fgcolor=(0, 0, 0))
_global_ns['figure'] = fig
for i, method in zip(count(), methods):
# if i in (3, 4):
# next_plot_fname(__file__, series='q3')
# print(i, "::", [line.shape[1] for line in all_lines[i]])
# # continue
vlab.clf()
_lshell_diff = [np.abs(s) for s in all_lshell_diffs[i]]
vlab.plot3d_lines(all_lines[i], scalars=_lshell_diff)
vlab.colorbar(title="Relative L-Shell Error (as fraction)")
vlab.title(method, size=0.5)
vlab.orientation_axes()
vlab.view(azimuth=40, elevation=140, distance=80.0,
focalpoint=[0, 0, 0])
vlab.savefig(next_plot_fname(__file__, series='q3'))
if args.show:
vlab.show()
except ImportError:
pass
# prevent weird xorg bad-instructions on tear down
if 'figure' in _global_ns and _global_ns['figure'] is not None:
from viscid.plot import vlab
vlab.mlab.close(_global_ns['figure'])
return 0