使用pyqtgraph绘制半导体晶圆图遇IndexError问题求助

问题描述 #

尝试用PyQtGraph的PColorMeshItem绘制由数千个不同颜色、坐标方块组成的半导体晶圆图，并实现悬停/点击事件，但运行时触发IndexError。

代码示例 #

'''
Demonstrates very basic use of PColorMeshItem
'''

import numpy as np
import pyqtgraph as pg

app = pg.mkQApp("PColorMesh Example")

## Create window with GraphicsView widget
win = pg.GraphicsLayoutWidget()
win.show()  ## show widget alone in its own window
win.setWindowTitle('pyqtgraph example: pColorMeshItem')
view = win.addViewBox()

## Create data
x_min = 0
x_max = 2
y_min = 0
y_max = 2
x = np.arange(x_min, x_max+2, 1, dtype=np.int16)
y = np.arange(y_min, y_max+2, 1, dtype=np.int16)
xmesh, ymesh = np.meshgrid(x, y, indexing='xy')
# init with all np.nan to hide all squares
z = np.full((y.size-1, x.size-1), np.nan)

# fill data in specific area
z[(1, 1, 1), (0, 1, 2)] = 1
z[(0, 1, 2), (1, 1, 1)] = 1

pcmi = pg.PColorMeshItem(xmesh, ymesh, z)
view.addItem(pcmi)

if __name__ == '__main__':
    pg.exec()

错误信息 #

File "/usr/local/lib/python3.9/site-packages/pyqtgraph/graphicsItems/PColorMeshItem.py", line 139, in __init__
    self.setData(*args)
  File "/usr/local/lib/python3.9/site-packages/pyqtgraph/graphicsItems/PColorMeshItem.py", line 258, in setData
    brushes = [lut[z] for z in norm[i].tolist()]
  File "/usr/local/lib/python3.9/site-packages/pyqtgraph/graphicsItems/PColorMeshItem.py", line 258, in <listcomp>
    brushes = [lut[z] for z in norm[i].tolist()]
IndexError: list index out of range

norm数组内容 #

[[-9223372036854775808 -9223372036854775808 -9223372036854775808]
 [-9223372036854775808 -9223372036854775808 -9223372036854775808]
 [-9223372036854775808 -9223372036854775808 -9223372036854775808]]

错误原因与修复 #

原因分析 #

错误源于z数组中混合了np.nan和整数1，PColorMeshItem处理时会将np.nan转换为极小整数（即norm数组中的-9223372036854775808），而默认颜色查找表（LUT）长度不足，导致索引越界。

修复步骤 #

1. 显式指定颜色映射范围：确保z的有效值落在颜色映射索引范围内，同时处理np.nan的显示。
2. 设置自定义颜色查找表（LUT）：或直接指定levels参数，让PyQtGraph自动生成适配的LUT。

修改后的代码示例：

'''
Fixed PColorMeshItem Example
'''

import numpy as np
import pyqtgraph as pg

app = pg.mkQApp("PColorMesh Example")

win = pg.GraphicsLayoutWidget()
win.show()
win.setWindowTitle('pyqtgraph example: pColorMeshItem')
view = win.addViewBox()

## Create data
x_min = 0
x_max = 2
y_min = 0
y_max = 2
x = np.arange(x_min, x_max+2, 1, dtype=np.int16)
y = np.arange(y_min, y_max+2, 1, dtype=np.int16)
xmesh, ymesh = np.meshgrid(x, y, indexing='xy')
z = np.full((y.size-1, x.size-1), np.nan)

# fill data in specific area
z[(1, 1, 1), (0, 1, 2)] = 1
z[(0, 1, 2), (1, 1, 1)] = 1

# 显式设置levels，指定z的有效值范围，同时让nan透明
pcmi = pg.PColorMeshItem(xmesh, ymesh, z, levels=[0, 2], antialias=False)
# 可选：自定义颜色映射
# cmap = pg.colormap.get('viridis')
# pcmi.setColorMap(cmap)

view.addItem(pcmi)

if __name__ == '__main__':
    pg.exec()

也可以将z中的np.nan替换为有效值范围外的数值（比如-1），再通过levels排除该值的显示，效果一致。

绘制半导体晶圆图的其他方案 #

• 使用ImageItem：如果晶圆方块是规则网格排列，ImageItem渲染效率更高，支持快速交互。只需将z数组作为图像数据传入，设置颜色映射即可，通过MouseEvents实现悬停/点击事件。
• 使用ScatterPlotItem：如果方块是非规则分布，可用ScatterPlotItem绘制正方形标记，通过sizes控制方块大小，brush设置颜色。这种方式灵活度高，适合非规则布局的晶圆图，交互事件易实现。
• 自定义GraphicsItem：针对高度定制化需求（如复杂交互逻辑、特殊渲染效果），可继承pyqtgraph.GraphicsObject自定义绘图项，在paint方法中直接绘制每个方块，完全控制渲染和交互行为。

内容的提问来源于stack exchange，提问作者nochenon