The colored analytic landscape of a complex function, f, is the graph of the function modulus or log-modulus, colored according to the argument of $f(z)$ at any z in the function domain.
Here we illustrate how defining this surface as an instance of the Plotly class plotly.graph_objects.Surface
, we can plot the contour lines as well as the surface projection onto a z-plane, that amounts to plotting the domain coloring of the function f.
We are using the cyclic HSV colorscale to color-encode the argument value. For details on HSV and domain coloring see this Jupyter notebook https://nbviewer.jupyter.org/github/empet/Math/blob/master/DomainColoring.ipynb.
import numpy as np
from numpy import pi
import plotly.graph_objs as go
import plotly.io as pio
pio.renderers.default ="notebook_connected"
def eval_modulus(f, re=(-2.5, 2.5), im=(-2.5, 2.5), N=50, log = False):
nrx = int(N * (re[1]-re[0]))
nry = int(N * (im[1]-im[0]))
x = np.linspace(re[0], re[1], nrx)
y = np.linspace(im[0], im[1], nry)
x, y = np.meshgrid(x, y)
z = x + 1j*y
w = f(z)
w[np.isinf(w)] = np.nan
if log:
modf = np.log(np.absolute(w))
else:
modf = np.absolute(w) # |f|
return x, y, np.angle(w), modf #np.angle(w) is the argument of f
HSV colorscale:
pl_hsv = [[0.0, 'rgb(0, 242, 242)'],
[0.083, 'rgb(0, 121, 242)'],
[0.167, 'rgb(0, 0, 242)'],
[0.25, 'rgb(121, 0, 242)'],
[0.333, 'rgb(242, 0, 242)'],
[0.417, 'rgb(242, 0, 121)'],
[0.5, 'rgb(242, 0, 0)'],
[0.583, 'rgb(242, 121, 0)'],
[0.667, 'rgb(242, 242, 0)'],
[0.75, 'rgb(121, 242, 0)'],
[0.833, 'rgb(0, 242, 0)'],
[0.917, 'rgb(0, 242, 121)'],
[1.0, 'rgb(0, 242, 242)']]
Define tickvals
and ticktext
for colorbar:
tickvals=[-np.pi, -2*np.pi/3, -np.pi/3, 0, np.pi/3, 2*np.pi/3, np.pi]
#define the above values as strings with pi-unicode
ticktext=['-\u03c0', '-2\u03c0/3', '-\u03c0/3', '0', '\u03c0/3', '2\u03c0/3', '\u03c0']
coloraxis_settings = dict(colorscale= pl_hsv,
colorbar_thickness=25,
colorbar_len=0.7,
colorbar_tickvals=tickvals,
colorbar_ticktext=ticktext,
colorbar_title='arg(f)')
def set_contours(min_mod, zrange_max, n=20, color = 'rgb(250, 250, 250)'):
return dict(start=min_mod,
end=zrange_max, highlight=True,
size=(zrange_max-min_mod)/n,
width=1.5, #contour line width
color= color,
project_z=True)
Let us plot the analytic landscape of the function f, defined below. f has a zero multiple of order 3, and three simple poles.
f = lambda z: z**3 / (z**3-1)
x, y, argf, modf = eval_modulus(f)
fig1 = go.Figure(go.Surface(x=x[0, :], y=y[:, 0], z=modf,
surfacecolor=argf, coloraxis='coloraxis'))
z_range = (-4, 6)
fig1.update_layout(title_text = '$\\text{Analytic landscape of the function}\: f(z)= \\displaystyle\\frac{z^3}{z^3-1}$',
title_x=0.5, font_family="Balto",
width=700, height=700,
coloraxis = coloraxis_settings,
scene_zaxis_range=z_range);
Add the surface projection onto the z-plane of equation z=z_range[0]
, and the contour lines of the plotted surface, and their projection, to illustrate the domain coloring plot of the function f(z) onto this z-plane:
min_mod = np.min(modf)
fig1.add_surface(x=x[0, :], y=y[:, 0], z= z_range[0]*np.ones(modf.shape),
surfacecolor=argf, colorscale= pl_hsv, showscale=False)
fig1.data[0].update(contours_z=dict(show=True, **set_contours(min_mod, z_range[1], n=28, color = 'rgb(250, 250, 250)')));
fig1.update_scenes(camera_eye_z=0.75);
fig1.show()
Finally let us plot the log modulus of $h(z)= e^{1/z}$:
h = lambda z: np.exp(1/z)
x, y, argh, modh = eval_modulus(h, log=True)
fig2 = go.Figure(go.Surface(x=x[0, :], y=y[:, 0], z=modh,
surfacecolor=argh,
coloraxis='coloraxis'))
z_range= (-6, 6)
fig2.update_layout(title_text = '$\\text{Analytic landscape of the function}\: f(z)= e^{1/z}$',
title_x=0.5, font_family="Balto",
width=700, height=700,
coloraxis = coloraxis_settings,
scene_zaxis_range=z_range);
min_mod = np.min(modh)
fig2.add_surface(x=x[0, :], y=y[:, 0], z= z_range[0]*np.ones(modh.shape),
surfacecolor=argh, colorscale= pl_hsv, showscale=False)
fig2.data[0].update(contours_z=dict(show=True,
start=z_range[0],
end=z_range[1], highlight=True,
size=(z_range[1]-z_range[0])/26,
width=1.5, #contour line width
color= 'rgb(250,250,250)',
project_z=True))
fig2.update_scenes(camera_eye_x=-1.55, camera_eye_y=1.55, camera_eye_z=0.6);
fig2.show()
Unlike the representation of modulus, $|f}$, the log-modulus, $log(|f|$, has negative values where $0<|f|<1$