Python matplotlib : How to tick and ticklabel an axis/object independently of the object's scale

Question

下面的代码在图中定义了一个矩形区域，并根据其中的颜色图绘制了一个颜色条。我已将颜色条的缩放比例更改为其原始值的立方根，但我希望刻度线和刻度标签在条形长度上保持线性，因此它对于我的目的来说是正确的。如果可能的话，我不妨在那里画一条直线，然后将其线性渐变。我该如何处理这个问题？ mpl 是 matplotlib，plt 是 pyplot。这个想法是为 choropleth(彩色世界地图)制作一个颜色条图例。

def draw_legend (clr_map):
    """ Draw color bar legend ... (inspired by: http://ramiro.org/notebook/basemap-choropleth/) """
    fig = plt.figure(figsize=(8,12))
    ax_legend = fig.add_axes([0.26, -0.02, 0.48, 0.016], zorder=3)
    grads = np.linspace(0.,1.,400)
    bins = np.linspace(0.,1.,10)
    scheme = [clr_map(i/400) for i in range(400)]
    cmap = mpl.colors.ListedColormap(scheme)
    cb = mpl.colorbar.ColorbarBase(ax_legend, cmap=cmap, ticks=bins, boundaries=grads**(1/3.), \
                               orientation='horizontal')   # Before I'd done cmap as sqrt(): boundaries=np.sqrt(grads)
    #cb.ax.set_xticks(bins)
    cb.ax.set_xticklabels([str(round(i, 1)) for i in bins], fontsize=10);

draw_legend (plt.cm.plasma)

Answer 1

要设置颜色条上的刻度，您可以使用cb.set_ticks(bins**(1/3.))。您也可以直接缩放颜色图 (clr_map((i/400.)**(1./3)))。

import matplotlib.pyplot as plt
import matplotlib.colors
import matplotlib.colorbar
import numpy as np

def draw_legend (clr_map):
    """ Draw a color bar legend 
        with a qubic root colormap
    """
    # Version 1, scale boundaries, set ticks to colorbar values
    fig = plt.figure(figsize=(6,4))
    ax_legend = fig.add_axes([0.26, 0.7, 0.48, 0.1], zorder=3)
    ax_legend.set_title("Version 1\nscale boundaries, set ticks to colorbar values")
    grads = np.linspace(0.,1.,400)
    bins = np.linspace(0.,1.,11)
    scheme = [clr_map(i/400.) for i in range(400)]
    cmap = matplotlib.colors.ListedColormap(scheme)
    cb = matplotlib.colorbar.ColorbarBase(ax_legend, cmap=cmap, ticks=bins, boundaries=grads**(1/3.), \
                               orientation='horizontal')   

    cb.set_ticks(bins**(1/3.))
    cb.ax.set_xticklabels(bins**(1/3.), fontsize=10, rotation =45, ha="center")
    cb.draw_all()

    # Version 2, scale colormap, set ticks to arbitrary values
    ax_legend2 = fig.add_axes([0.26, 0.27, 0.48, 0.1], zorder=3)
    ax_legend2.set_title("Version 2\nscale colormap, set ticks to arbitrary values")
    grads = np.linspace(0.,1.,400)
    bins = np.linspace(0.,1.,11)
    scheme = [clr_map((i/400.)**(1./3)) for i in range(400)]
    cmap = matplotlib.colors.ListedColormap(scheme)
    cb = matplotlib.colorbar.ColorbarBase(ax_legend2, cmap=cmap, ticks=bins, 
                               orientation='horizontal')   
    cb.set_ticks(bins)
    cb.draw_all()


draw_legend (plt.cm.jet)
plt.savefig(__file__+".png")
plt.show()