Plotly导出图像时Colorbar异常变为单色的解决方法

尝试用Plotly制作3D锥图，使用仅改变颜色透明度的自定义colorscale，预览窗口显示正常：

但通过快照图标导出图像后，色阶变为单色：

相关代码如下：

import numpy as np
import plotly.graph_objects as go
from emknot import *

x = np.linspace(-3, 3, 7)
y = np.linspace(-3, 3, 7)
z = np.linspace(-3, 3, 7)


cones = []

for i in x:
    for j in y:
        for k in z: 
            print(i)
            u = [globule([i, j, k],k=1)[0]]
            v = [globule([i, j, k],k=1)[1]]
            w = [globule([i, j, k],k=1)[2]]

            cones.append(go.Cone(
            x=[i],
            y=[j],
            z=[k],
            u=u,
            v=v,
            w=w,
            cmin=0,
            cmax=1.5,
            colorscale=[[0,'rgba(255,255,255, 0)'],[1,'rgba(0,0,255,1)']],
            sizemode="absolute",
            sizeref=0.6,
            anchor="center")) 


layout = go.Layout(title='EM-globule')
fig = go.Figure(data=cones, layout=layout)

fig.show()

这个问题源于Plotly浏览器快照功能对RGBA透明度渐变的支持局限，导出时会丢失透明度层级信息。可以通过以下两种方式解决：

方法1：手动控制锥颜色透明度，自定义色阶 #

放弃自动生成的色阶，直接根据每个锥的数值计算透明度，固定颜色通道，再手动添加模拟的色阶条：

import numpy as np
import plotly.graph_objects as go
from emknot import *

x = np.linspace(-3, 3, 7)
y = np.linspace(-3, 3, 7)
z = np.linspace(-3, 3, 7)

cones = []
cmin = 0
cmax = 1.5

for i in x:
    for j in y:
        for k in z: 
            vec = globule([i, j, k], k=1)
            u, v, w = [vec[0]], [vec[1]], [vec[2]]
            # 用向量模长作为颜色映射值
            val = np.linalg.norm(vec)
            # 计算对应透明度
            alpha = np.clip((val - cmin)/(cmax - cmin), 0, 1)
            cones.append(go.Cone(
                x=[i],
                y=[j],
                z=[k],
                u=u,
                v=v,
                w=w,
                color=f'rgba(0,0,255, {alpha})',
                sizemode="absolute",
                sizeref=0.6,
                anchor="center",
                showscale=False
            )) 

# 构建图形并手动添加色阶
fig = go.Figure(data=cones, layout=go.Layout(title='EM-globule'))
# 色阶外框
fig.add_shape(
    type="rect",
    x0=0.9, x1=0.92, y0=0.1, y1=0.9,
    fillcolor="rgba(0,0,255,0)",
    line_color="black",
    xref="paper", yref="paper"
)
# 渐变填充条
for i in range(10):
    alpha = i/9
    fig.add_shape(
        type="rect",
        x0=0.9, x1=0.92, y0=0.1 + i*0.08, y1=0.1 + (i+1)*0.08,
        fillcolor=f"rgba(0,0,255, {alpha})",
        line_width=0,
        xref="paper", yref="paper"
)
# 色阶数值标签
fig.add_annotation(x=0.95, y=0.9, text=f"{cmax}", xref="paper", yref="paper", showarrow=False)
fig.add_annotation(x=0.95, y=0.1, text=f"{cmin}", xref="paper", yref="paper", showarrow=False)

fig.show()

方法2：使用Plotly专业导出方法 #

Plotly的浏览器快照依赖前端渲染，而write_image基于Kaleido后端，对透明度支持更完善：

1. 先安装依赖：

pip install kaleido

2. 修改代码末尾，用write_image导出：

# 保留原有代码，替换fig.show()部分
fig.show()
# 导出为PNG格式，效果与预览一致
fig.write_image("em_globule.png")

内容的提问来源于stack exchange，提问作者Aceman