Description
Bug summary
I use LinearSegmentedColormap.from_list to create custom colormaps. A very simple one, which just assigns two RGB values to the two end points, does not work any more after update to matplotlib 3.8.0.
Code for reproduction
from matplotlib import colors
val_col = [(0.0, (0.95, 0.0, 0.0)), (1.0, (0.95, 0.95, 0.95))]
cm = colors.LinearSegmentedColormap.from_list('test', val_col)Actual outcome
File ~/packages/macports/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/matplotlib/colors.py:1067, in LinearSegmentedColormap.from_list(name, colors, N, gamma)
1063 vals = np.linspace(0, 1, len(colors))
1065 r, g, b, a = to_rgba_array(colors).T
1066 cdict = {
-> 1067 "red": np.column_stack([vals, r, r]),
1068 "green": np.column_stack([vals, g, g]),
1069 "blue": np.column_stack([vals, b, b]),
1070 "alpha": np.column_stack([vals, a, a]),
1071 }
1073 return LinearSegmentedColormap(name, cdict, N, gamma)
File ~/packages/macports/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/numpy/lib/shape_base.py:652, in column_stack(tup)
650 arr = array(arr, copy=False, subok=True, ndmin=2).T
651 arrays.append(arr)
--> 652 return _nx.concatenate(arrays, 1)
ValueError: all the input array dimensions except for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 2 and the array at index 1 has size 3
Expected outcome
This should construct a new colormap going smoothly from red to white. This used to work until I recently updated matplotlib to 3.8.0.
Additional information
The issue seems to be that the routine to_rgba_array takes a length-2 object and returns r, g and b that are all length 3. Specifically, the line
r, g, b, a = to_rgba_array(colors).Ttakes colors, which is ((0.95, 0.0, 0.0), (0.95, 0.95, 0.95)), and returns r as [0.95 0.95 0.95], and so on. It's obvious that np.column_stack([vals, r, r]) should fail in that case, because vals is length 2, and r is length 3. This used to work in an earlier version of matplotlib, so I suppose to_rgba_array has changed recently.
Operating system
OS/X
Matplotlib Version
3.8.0
Matplotlib Backend
MacOSX
Python version
3.11.6
Jupyter version
No response
Installation
Linux package manager