python 随机颜色

python 随机颜色

从调色板种均匀取出一些颜色，颜色不会重复，颜色数量可以随机
参考：https://stackoverflow.com/questions/55469432/is-there-a-similar-color-palette-to-tab20c-with-bigger-number-of-colors
颜色参考：https://matplotlib.org/stable/tutorials/colors/colormaps.html

def random_colormap(N: int,cmaps_='gist_ncar',show_=False):
    # 从颜色图（梯度多）中取N个
    # test_cmaps = ['gist_rainbow', 'nipy_spectral', 'gist_ncar']
    cmap = matplotlib.colors.ListedColormap(plt.get_cmap(cmaps_)(np.linspace(0, 1, N)))
    if show_:
        gradient = np.linspace(0, 1, 256)
        gradient = np.vstack((gradient, gradient))
        fig, ax = plt.subplots(1, 1, figsize=(5, 1))
        ax.imshow(gradient, aspect='auto', cmap=cmap)
        plt.show()
    return cmap

从不同的颜色图中取颜色并合并,每个base中颜色梯度为20，也就是最多取60个颜色
base_cmaps = [‘tab20’,‘tab20b’,‘tab20c’]可以添加其他palette，但是要注意颜色梯度数量

def random_colormap2(N: int,base_cmaps=['gist_ncar'],show_=False):
    # 从不同的颜色图中取颜色并合并,每个base中颜色梯度为20，也就是最多60个颜色
    base_cmaps = ['tab20','tab20b','tab20c']
    N //= len(base_cmaps)
    # N += 1
    # base_cmaps = ['Greys', 'Purples', 'Reds', 'Blues', 'Oranges', 'Greens']
    # segmented_cmaps = [matplotlib.colors.ListedColormap(plt.get_cmap(t)(np.linspace(0,1,N))) for t in test_cmaps]
    colors = np.concatenate([plt.get_cmap(name)(np.linspace(0, 1, N)) for name in base_cmaps])
    cmap = matplotlib.colors.ListedColormap(colors)
    if show_:
        gradient = np.linspace(0, 1, 256)
        gradient = np.vstack((gradient, gradient))
        fig, ax = plt.subplots(1,1,figsize=(5,1))
        ax.imshow(gradient, aspect='auto', cmap=cmap)
        plt.show()
    return cmap

• 使用，返回值是cmap

  random_colormap(40, cmaps_='gist_ncar', show_=True)
  cmaps = random_colormap2(60, base_cmaps=['gist_ncar'], show_=True)
  

如果颜色太多，推荐使用第一种，毕竟颜色梯度变化是连续的