0

I am trying to set up a 3D scatter plot animation, following the example here: Matplotlib 3D scatter animations

I put together the following code: 

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as animation
from mpl_toolkits.mplot3d import Axes3D

def terminal_w(r):
    """calculation of droplet terminal fallspeed"""
    w=X1*r**2
    return(w)

class drop():
    """class of a drop"""
    def __init__(self,xpos,ypos,zpos,rad):
        self.x,self.y,self.z=xpos,ypos,zpos # position
        self.r=rad     # radius 
        self.w=terminal_w(rad) # velocity  


def main():
    # dimensions of the domain
    global xsize,ysize,zsize
    global X1,X2,X3
    global dt # timestep
    global rho_l

    xsize,ysize,zsize=1.,1.,100.    # domain size in meters
    X1,X2,X3=1.2e8,8e3,250. # terminal velocity coeffs
    dt=10.0
    rho_l=1000. # density of liquid water

    #
    # Student exercise 1: change radii and liquid water amount:
    #
    liq=1.e-3 # cloud liq water content in kg/m**3    
    rsmall=5.e-6 # microns
    rlarge=20.e-6

    # L=N*rho_l*4/3 pi r^3
    ndrop=10 # total number of drops
    # initial number of large drops with rlarge
    # not used if distrbution of drop sizes assumed
    nlarge=1  

    # initial random positions:
    # could use a dictionary, but don't want to lose numpy advantage?
    dropx=np.random.uniform(low=0,high=xsize,size=ndrop)
    dropy=np.random.uniform(low=0,high=ysize,size=ndrop)
    dropz=np.random.uniform(low=0,high=zsize,size=ndrop)

    # one large drop falling through a cloud of small drops
    dropr=np.full(ndrop,fill_value=rsmall)
    dropr[-nlarge:]=rlarge

    # Student exercise 2: change distribution: 
    # can insert code here to simply set a distribution of radii:
    # set arrange from lognormal distribution, 
    # dropr=np.random.DIST(moments)

    # initial drop conditions
    drops=drop(dropx,dropy,dropz,dropr)

    # set up plot window
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')

    ax.set_xlim(0,xsize)
    ax.set_ylim(0,ysize)
    ax.set_zlim(0,zsize)

    sc3d=ax.scatter(drops.x,drops.y,drops.z,c='blue',marker='o')
    title = ax.set_title('3D Test')

    def animate(i):
        timestep(drops)
        print(drops.z[0])
        sc3d._offsets3d=(drops.x,drops.y,drops.z)  # update the data
        title.set_text('3D Test, time={}'.format(i))

    ani = animation.FuncAnimation(fig, animate, 
                                  interval=25, 
                                  blit=False)

    plt.show()

def timestep(drops):
    #print(drops.w)
    drops.z-=dt*drops.w
    #print(drops.z)


main()

which produces a 3D scatter plot, but only animates ONE of the markers, even though all entries are having their z coordinates updated (when I uncheck the print statements).

Even more bizarrely, if I comment out the statement 

title.set_text('3D Test, time={}'.format(i))

in the animate function, none of the drops are animated. I don't understand why adding the title statement allows one marker to animate, and also I can't see how my code differs from the example which functions for me and animates all points...

ClimateUnboxed
  • 7,106
  • 3
  • 41
  • 86
  • The code as it stands animates correctly. I suppose you are tricked by the fact that one of the drops is ***much faster*** than then others? – ImportanceOfBeingErnest Nov 29 '19 at 14:57
  • I'm not sure - when I put the return statement in I see all drops move, albeit with the small radiii ones much slower. I still haven't fully understood yet, I'm using this animate feature for the first time. – ClimateUnboxed Nov 29 '19 at 21:45

1 Answers1

1

According to the docs, your animate function signature should be def func(frame, *fargs) -> iterable_of_artists. However, you are not returning an iterable of artists. You aren't returning anything!

EDIT: Just to be clear, your function signature is func(frame) -> None. Oh, and don't use globals, they are evil.

FiddleStix
  • 3,016
  • 20
  • 21