在热图中使反向对角线变白

Question

我正在尝试做一些如下图所示的图片，

只需设置反向对角线，就会留下白色。我无法将它们设置为白色。图表采用整数值，我不知道白色对应的整数值。

感谢！

编辑：

这是代码;

import math
from matplotlib import pyplot as plt
from matplotlib import cm as cm
import pylab
import numpy as np
from matplotlib.collections import LineCollection

class HeatMap:
    def __init__(self, selectedLines):
        self.selectedLines = selectedLines


    def getHeapMap(self):
        figure = plt.figure()

        if len(self.selectedLines) != 0:

            self.map = self.createTestMapData(len(self.selectedLines),                len(self.selectedLines))


            maxValueInMap = self.findMaxValueInMap(self.map)

            x = np.arange(maxValueInMap + 1)
            ys = [x + i for i in x]
            ax = figure.add_subplot(111)
            ax.imshow(self.map, cmap=cm.jet, interpolation='nearest')

            '''
            Left side label of the chart is created according to selected values
            from a checkbox group.
            '''
            leftSideLabelSize = len(self.selectedLines)
            sideLabels = []
            for line in self.selectedLines:
                sideLabels.append(line.text())
            pos = np.arange(leftSideLabelSize)
            '''
            Left side labels are set with the code below.
            '''
            pylab.yticks(pos, sideLabels)
            plt.xticks(pos, sideLabels)
            self.numrows, self.numcols = self.map.shape
            ax.format_coord = self.format_coord

            line_segments = LineCollection([zip(x, y) for y in ys],
                linewidths=(0.5, 3, 1.5, 2),
                linestyles='solid')
            line_segments.set_array(x)
            axcb = figure.colorbar(line_segments)

        return figure

    def format_coord(self, x, y):
        col = int(x + 0.5)
        row = int(y + 0.5)
        if col >= 0 and col < self.numcols and row >= 0 and row < self.numrows:
            z = self.map[row, col]
            return 'x=%1.4f, y=%1.4f, z=%1.4f' % (x, y, z)
        else:
            return 'x=%1.4f, y=%1.4f' % (x, y)

    def createTestMapData(self, xSize, ySize):
        resultMap = 10 * np.random.rand(xSize, ySize)
        #Setting reverse diagonal is here. Now it is set with zero but it gives blue.
        # I want it to be set as white
        for index in range(0, int(math.sqrt(resultMap.size))):
            resultMap[index][((math.sqrt(resultMap.size) - 1) - index )] = 0 
        return  resultMap

    def findMaxValueInMap(self, map):
        return np.amax(map)

此时会随机生成这些值。上面的代码给出了一个gui之类的;

Answer 1

您可以制作自己的色彩映射表，也可以调整现有的色彩映射表：）

以下是上述情节的代码，注释中有解释：

import matplotlib
from pylab import *
import numpy as np

#Create test data with zero valued diagonal:
data = np.random.random_sample((25, 25))
rows, cols = np.indices((25,25))
data[np.diag(rows, k=0), np.diag(cols, k=0)] = 0

#Create new colormap, with white for zero 
#(can also take RGB values, like (255,255,255):
colors = [('white')] + [(cm.jet(i)) for i in xrange(1,256)]
new_map = matplotlib.colors.LinearSegmentedColormap.from_list('new_map', colors, N=256)

pcolor(data, cmap=new_map)
colorbar()
savefig('map.png')
show()

或者，您可以屏蔽数据，并设置蒙版颜色：

#Create test data:
data = np.random.random_sample((25, 25))
#Create a diagonal mask:
mask = np.diag(np.ones(25))
#Apply mask to data:
masked_data = ma.masked_array(data, mask)
#Set mask color to white:
cm.jet.set_bad(color='white', alpha=None)
#for this to work we use pcolormesh instead of pcolor:
pcolormesh(masked_data, cmap=cm.jet)
colorbar()
show()

这会产生基本相同的结果，但可以更好地满足您的需求，因为您可以将任何单元格设置为白色，并且白色不会显示在颜色条上（请参见上面颜色条的底部）：

Answer 2

色彩映射由cmap中的ax.imshow()参数定义。您使用了jet色彩映射，因此您拥有cmap=cm.jet，这只是matplotlib中众多built-in color maps中的一个。您可以选择一个或定义适合自己口味的自己。