Matplotlib:生成具有不同比例和反向比例的多个双轴 [英] Matplotlib: generate multiple twin axes with different and inverted scale

问题描述

我想在两个 x 和 y 轴上绘制一个数据系列，以便有 4 个不同的轴.首先是 x(以 eV 为单位的能量)与 y(归一化计数)轴，然后是 x(与能量成反比的波长)与 y(计数)轴.我的代码是:

I want to plot a data series on two x and y axes in order to have 4 different axes. First the x (energy in eV) vs. the y (normalized counts) axis and then x (wavelength which is inversely related to energy) vs. y (counts) axis. My code for this is:

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.mlab as mlab
from scipy.constants import h, c, e


def E(wavelength):
   return (h*c)/(wavelength*e)

wavelen = np.linspace(800e-9,1600e-9,200)
E_eV = E(wavelen)
loc, scale = 950e-9, 3.0
counts = mlab.normpdf(wavelen,950e-9,100e-9)/100
counts_norm = counts/10000


fig, ax  = plt.subplots()

ax1 = ax
ax2 = ax.twinx()
ax3 = ax.twiny()

plt.ticklabel_format(style='sci', scilimits=(0,0))

ax1.plot(E_eV, counts_norm)
ax1.set_xlim(E(1600e-9),E(800e-9))
ax1.set_ylabel('normalized counts')
ax1.set_xlabel('energy (eV)')
ax2.plot(E_eV, counts)
ax2.set_xlim(E(1600e-9),E(800e-9))
ax2.set_ylabel('counts')
ax3.plot(wavelen*1e9, counts_norm)
ax3.set_xlim(1600,800)
ax3.set_xlabel('wavelength (nm)')
ax3.ticklabel_format(style='plain')


plt.tight_layout()
plt.show()

如您所见，曲线的缩放比例不正确，因此它们重叠并在x方向上具有相同的尺寸.你能帮我如何为顶部的 x(波长)轴设置正确的参数吗?

As you can see the curves are not scaled in the right way so that they overlap and have the same dimensions in x-direction. Can you help me how to set the right parameters for the x (wavelength) axis at the top?

推荐答案

我建议只在主轴上绘图，然后同步双轴的标签.我编辑了您的示例，以展示如何针对静态图完成此操作.

I recommend plotting only on your primary axes and then synchronizing the labels of the twin axes. I edited your example to show how this could be accomplished for a static plot.

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.mlab as mlab
from scipy.constants import h, c, e

def E(wavelength):
    return (h*c)/(wavelength*e)
def getWaveLength(energy):
    return (h*c)/(energy*e)
def getCounts(normcounts):
    return normcounts*1000

wavelen = np.linspace(800e-9,1600e-9,200)
E_eV = E(wavelen)
loc, scale = 950e-9, 3.0
counts = mlab.normpdf(wavelen,950e-9,100e-9)/100
counts_norm = counts/10000

fig, ax1  = plt.subplots()

ax2 = ax1.twinx()
ax3 = ax1.twiny()

plt.ticklabel_format(style='sci', scilimits=(0,0))

ax1.plot(E_eV, counts_norm)
ax1.set_xlim(E(1600e-9),E(800e-9))
ax1.set_ylabel('normalized counts')
ax1.set_xlabel('energy (eV)')
ax2.set_ylabel('counts')
ax3.set_xlabel('wavelength (nm)')
ax3.ticklabel_format(style='plain')

# get the primary axis x tick locations in plot units
xtickloc = ax1.get_xticks() 
# set the second axis ticks to the same locations
ax3.set_xticks(xtickloc)
# calculate new values for the second axis tick labels, format them, and set them
x2labels = ['{:.3g}'.format(x) for x in getWaveLength(xtickloc)]
ax3.set_xticklabels(x2labels)
# force the bounds to be the same
ax3.set_xlim(ax1.get_xlim()) 

#same for y
ytickloc = ax1.get_yticks()
ax2.set_yticks(ytickloc)
ax2.set_yticklabels([str(int(y)) for y in getCounts(ytickloc)])
ax2.set_ylim(ax1.get_ylim())

plt.tight_layout()
plt.show()

这篇关于Matplotlib:生成具有不同比例和反向比例的多个双轴的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！