動作環境

GeForce GTX 1070 (8GB)
ASRock Z170M Pro4S [Intel Z170chipset]
Ubuntu 16.04 LTS desktop amd64
TensorFlow v1.2.1
cuDNN v5.1 for Linux
CUDA v8.0
Python 3.5.2
IPython 6.0.0 -- An enhanced Interactive Python.
gcc (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609
GNU bash, version 4.3.48(1)-release (x86_64-pc-linux-gnu)
scipy v0.19.1
geopandas v0.3.0
MATLAB R2017b (Home Edition)
ADDA v.1.3b6

概要

20万個の粒子の描画をJupyter + Matplotlibで実装中である。

処理の大半はdunder mainを使わずに実装していた。

[処理1]
[処理2]
...

Jupyter上でのデバッグ効率化のため、dunder main使用に変更してみた。

def main():
    [処理1]
    [処理2]

if __name__ == '__main__':
    main()

その結果、描画がされなくなった。

code

showParticleShape_180107-Copy1.ipynb

import matplotlib.pyplot as plt
from matplotlib import cm, colors
from mpl_toolkits.mplot3d import Axes3D
from pylab import rcParams
import numpy as np
import time
import contextlib


"""
*** Copied for test 2018/01/20 ***

"""

"""
v0.7 Jan. 08, 2018
  - save three view images (top, front, side)
v0.6 Jan. 08, 2018
  - print progress of the processing using the contextmanager
     - add using_oneline_print()
v0.5 Jan. 07, 2018
  - rename to [showParticleShape_18017.ipynb]
    + was [showChebyshev_170910.ipynb]
v0.4 Sep. 10, 2017
  - read ADDA file [IntField-Y]
v0.3 Sep. 10, 2017
  - set colors in X direction
  - increase number of spheres to 3000
v0.2 Sep. 10, 2017
  - show 8 spheres
  - lower the resolution of the sphere (from 100j to 6j)
v0.1 Sep. 10, 2017
  - show 2 spheres
"""

# coding rule: PEP8

rcParams['figure.figsize'] = 15, 10


# reference
# https://stackoverflow.com/questions/31768031/plotting-points-on-the-surface-of-a-sphere-in-pythons-matplotlib


# Reference:
# countdown.py at
# https://qiita.com/QUANON/items/c5868b6c65f8062f5876
# by @QUANON
@contextlib.contextmanager
def oneline_print():
    import sys
    original_write = sys.stdout.write  # keep the original function

    def overwrite(text):
        # delete from the cursor position to the end of line
        original_write('\033[K')
        original_write('\r')  # move to the beginning of line
        # remove new line at the end from the text and print()
        original_write(text.rstrip('\n'))

    sys.stdout.write = overwrite  # replace the function

    try:
        yield
    finally:
        sys.stdout.write = original_write  # set the original function
        print('*** [sys.stdout.write] is now the original')


def plot_spheres(xps, yps, zps):
    DBG = 0
    for idx, elem in enumerate(zip(xps, yps, zps)):
        DBG += 1
        if DBG > 300:
            break
        axp, ayp, azp = elem
        print(axp, ayp, azp)
        #if idx % 10 == 0:
        #    # print progress
        #    msg = "%.2f " % (idx / len(xps) * 100) + "%"
        #    print(msg)
        dx = x + axp
        dy = y + ayp
        dz = z + azp
        ax.plot_surface(
            dx, dy, dz,  rstride=1, cstride=1, color='c',
            alpha=1.0, linewidth=0,
            # 50: arbitrary chosen to set colors
            facecolors=plt.cm.Set2((dx - 0) / (50 - 0)))

def main():
    start_time = time.time()

    # Create a sphere
    r = 1
    pi = np.pi
    cos = np.cos
    sin = np.sin
    phi, theta = np.mgrid[0.0:pi:6j, 0.0:2.0*pi:6j]
    x = r*sin(phi)*cos(theta)
    y = r*sin(phi)*sin(theta)
    z = r*cos(phi)


    # read from ADDA file
    INPFILE = 'IntField-Y'
    SHIFT_POS = 10.0
    dat = np.genfromtxt(INPFILE, delimiter=' ', skip_header=1)
    xpar, ypar, zpar = [], [], []
    for elem in dat:
        axp, ayp, azp = elem[:3]
        # 3.0: arbitrary chosen to adjust sphere positions
        xpar += [(axp + SHIFT_POS) * 3.0]
        ypar += [(ayp + SHIFT_POS) * 3.0]
        zpar += [(azp + SHIFT_POS) * 3.0]

    # Set colours and render
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')

    SKIP_NUM = 2
    xp = xpar[::SKIP_NUM]
    yp = ypar[::SKIP_NUM]
    zp = zpar[::SKIP_NUM]

    with oneline_print():
        plot_spheres(xp, yp, zp)
 
    ax.set_xlim([0, 50])
    ax.set_ylim([0, 50])
    ax.set_zlim([0, 50])
    ax.set_aspect("equal")
    ax.set_xlabel("X")
    ax.set_ylabel("Y")
    ax.set_zlabel("Z")
    plt.tight_layout()

    duration = time.time() - start_time
    print("start showing plot. elapsed: %.3f sec" % duration)

    # output three-view-images
    ax.view_init(elev=90, azim=0)
    plt.savefig("img_top.png")
    ax.view_init(elev=0, azim=90)
    plt.savefig("img_side.png")
    ax.view_init(elev=0, azim=0)
    plt.savefig("img_front.png")

    # show in Jupyter
    ax.view_init(elev=30, azim=-60)
    plt.show()

    duration = time.time() - start_time
    print("finish showing plot. elapsed: %.3f sec" % duration)

if __name__ == '__main__':
    main()

結果

注意: 上記実行にはファイルが必要です。

dunder mainなし

dunder main実行

備考

dunder mainというよりは、処理を関数にして、関数をコールした時に発生する症状のようだ。

def main():
   [処理1]
   [処理2]

main()

カーネルのRestart and output clear

カーネルを再起動し、出力をクリアした。
コードのいくつかがエラーになった。

キャッシュに基づき動作していたのかもしれない。

原因と対処

以下のようになっていた。

関数内でx,y,zを使っていた
x,y,zは関数外でグローバル変数として定義されていた
x,y,zの定義はmain()に移された(今回の変更)
x,y,zのキャッシュが使われ描画
- 描画されない

Matplotlibはコード変更直後に描画が時々されない経験がある。そういう場合はRun Cellsを実行すれば描画が正常になる場合がある。
今回は描画がされない状態のキャッシュが残り、そのキャッシュを使い続けることで、いつまでも描画されない状態になっていたのかもしれない。

対処

問題に気づくようにキャッシュをクリアした方が良いだろう

Kernel > Restart & Clear Output
など

code

showParticleShape_180107-Copy1.ipynb

import matplotlib.pyplot as plt
from matplotlib import cm, colors
from mpl_toolkits.mplot3d import Axes3D
from pylab import rcParams
import numpy as np
import time
import contextlib


"""
*** Copied for test 2018/01/20 ***

"""

"""
v0.7 Jan. 08, 2018
  - save three view images (top, front, side)
v0.6 Jan. 08, 2018
  - print progress of the processing using the contextmanager
     - add using_oneline_print()
v0.5 Jan. 07, 2018
  - rename to [showParticleShape_18017.ipynb]
    + was [showChebyshev_170910.ipynb]
v0.4 Sep. 10, 2017
  - read ADDA file [IntField-Y]
v0.3 Sep. 10, 2017
  - set colors in X direction
  - increase number of spheres to 3000
v0.2 Sep. 10, 2017
  - show 8 spheres
  - lower the resolution of the sphere (from 100j to 6j)
v0.1 Sep. 10, 2017
  - show 2 spheres
"""

# coding rule: PEP8

rcParams['figure.figsize'] = 15, 10


# reference
# https://stackoverflow.com/questions/31768031/plotting-points-on-the-surface-of-a-sphere-in-pythons-matplotlib


# Reference:
# countdown.py at
# https://qiita.com/QUANON/items/c5868b6c65f8062f5876
# by @QUANON
@contextlib.contextmanager
def oneline_print():
    import sys
    original_write = sys.stdout.write  # keep the original function

    def overwrite(text):
        # delete from the cursor position to the end of line
        original_write('\033[K')
        original_write('\r')  # move to the beginning of line
        # remove new line at the end from the text and print()
        original_write(text.rstrip('\n'))

    sys.stdout.write = overwrite  # replace the function

    try:
        yield
    finally:
        sys.stdout.write = original_write  # set the original function
        print('*** [sys.stdout.write] is now the original')


def plot_spheres(ax, cx, cy, cz, xps, yps, zps):
    DBG = 0
    for idx, elem in enumerate(zip(xps, yps, zps)):
        DBG += 1
        if DBG > 300:
            break
        axp, ayp, azp = elem
        #print(axp, ayp, azp)
        if idx % 10 == 0:
            # print progress
            msg = "%.2f " % (idx / len(xps) * 100) + "%"
            print(msg)
        dx = cx + axp
        dy = cy + ayp
        dz = cz + azp
        ax.plot_surface(
            dx, dy, dz,  rstride=1, cstride=1, color='c',
            alpha=1.0, linewidth=0,
            # 50: arbitrary chosen to set colors
            facecolors=plt.cm.Set2((dx - 0) / (50 - 0)))


def main():
    # Create a sphere
    r = 1
    pi = np.pi
    cos = np.cos
    sin = np.sin
    phi, theta = np.mgrid[0.0:pi:6j, 0.0:2.0*pi:6j]
    cx = r*sin(phi)*cos(theta)
    cy = r*sin(phi)*sin(theta)
    cz = r*cos(phi)

    #def main():
    start_time = time.time()

    # read from ADDA file
    INPFILE = 'IntField-Y'
    SHIFT_POS = 10.0
    dat = np.genfromtxt(INPFILE, delimiter=' ', skip_header=1)
    xpar, ypar, zpar = [], [], []
    for elem in dat:
        axp, ayp, azp = elem[:3]
        # 3.0: arbitrary chosen to adjust sphere positions
        xpar += [(axp + SHIFT_POS) * 3.0]
        ypar += [(ayp + SHIFT_POS) * 3.0]
        zpar += [(azp + SHIFT_POS) * 3.0]

    # Set colours and render
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')

    SKIP_NUM = 2
    xp = xpar[::SKIP_NUM]
    yp = ypar[::SKIP_NUM]
    zp = zpar[::SKIP_NUM]

    with oneline_print():
        plot_spheres(ax, cx, cy, cz, xp, yp, zp)

    ax.set_xlim([0, 50])
    ax.set_ylim([0, 50])
    ax.set_zlim([0, 50])
    ax.set_aspect("equal")
    ax.set_xlabel("X")
    ax.set_ylabel("Y")
    ax.set_zlabel("Z")
    plt.tight_layout()

    duration = time.time() - start_time
    print("start showing plot. elapsed: %.3f sec" % duration)

    # output three-view-images
    ax.view_init(elev=90, azim=0)
    plt.savefig("img_top.png")
    ax.view_init(elev=0, azim=90)
    plt.savefig("img_side.png")
    ax.view_init(elev=0, azim=0)
    plt.savefig("img_front.png")

    # show in Jupyter
    ax.view_init(elev=30, azim=-60)
    plt.show()

    duration = time.time() - start_time
    print("finish showing plot. elapsed: %.3f sec" % duration)

if __name__ == '__main__':
    main()

link

dunder mainの表記のもとになった記事。

Pythonの__init__なんて読むの問題