Bug summary
Minimal test case using LinearSegmentedColormap extracted from https://github.com/pyani-plus/pyani-plus/blob/v1.0.1/pyani_plus/plot_run.py which worked on macOS and Linux with eg matplotlib==3.10.8 but breaks on matplotlib-3.11.0 with:
ValueError: the values passed in the (value, color) pairs must increase monotonically from 0 to 1.
The release notes API changes do not mention this, https://matplotlib.org/stable/api/prev_api_changes/api_changes_3.11.0.html
Code for reproduction
from matplotlib import colormaps, __version__
from matplotlib.colors import LinearSegmentedColormap
print(f"matplotlib {__version__}")
GREY = (0.7, 0.7, 0.7)
DULL_BLUE = (0.137, 0.412, 0.737)
WHITE = (1.0, 1.0, 1.0)
DULL_RED = (0.659, 0.216, 0.231)
colormaps.register(
LinearSegmentedColormap.from_list(
"spbnd_BuRd", # species boundary - blue to red
(
(0.00, GREY), # 0% grey
(0.80, GREY), # 80% grey (not meaningful)
(0.80, DULL_BLUE), # 80% blue
(0.95, WHITE), # 95% white (species boundary)
(1.00, DULL_RED), # 100% red
),
)
)
print("Done")Actual outcome
It now fails:
❯ pip install matplotlib==3.11.0
Collecting matplotlib==3.11.0
...
Successfully installed matplotlib-3.11.0
❯ python test.py
matplotlib 3.11.0
Traceback (most recent call last):
File "/Users/peterjc/miniforge3/envs/pyani-plus_py312/lib/python3.12/site-packages/matplotlib/colors.py", line 1218, in from_list
r, g, b, a = to_rgba_array(colors).T
^^^^^^^^^^^^^^^^^^^^^
File "/Users/peterjc/miniforge3/envs/pyani-plus_py312/lib/python3.12/site-packages/matplotlib/colors.py", line 532, in to_rgba_array
rgba = np.array([to_rgba(cc) for cc in c])
^^^^^^^^^^^
File "/Users/peterjc/miniforge3/envs/pyani-plus_py312/lib/python3.12/site-packages/matplotlib/colors.py", line 345, in to_rgba
rgba = _to_rgba_no_colorcycle(c, alpha)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/peterjc/miniforge3/envs/pyani-plus_py312/lib/python3.12/site-packages/matplotlib/colors.py", line 364, in _to_rgba_no_colorcycle
if alpha is not None and not 0 <= alpha <= 1:
^^^^^^^^^^^^^^^
TypeError: '<=' not supported between instances of 'int' and 'tuple'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/private/tmp/test.py", line 12, in <module>
LinearSegmentedColormap.from_list(
File "/Users/peterjc/miniforge3/envs/pyani-plus_py312/lib/python3.12/site-packages/matplotlib/colors.py", line 1229, in from_list
raise ValueError(
ValueError: the values passed in the (value, color) pairs must increase monotonically from 0 to 1.
Expected outcome
This used to work, eg
❯ pip install matplotlib==3.10.8
Collecting matplotlib==3.10.8
...
Successfully installed matplotlib-3.10.8
❯ python test.py
matplotlib 3.10.8
Done
Additional information
No response
Operating system
No response
Matplotlib Version
3.11.0
Matplotlib Backend
macosx (or Linux tested via CI)
Python version
3.12.10 (and Python 3.11 and 3.13 via CI)
Jupyter version
No response
Installation
pip
Bug summary
Minimal test case using
LinearSegmentedColormapextracted from https://github.com/pyani-plus/pyani-plus/blob/v1.0.1/pyani_plus/plot_run.py which worked on macOS and Linux with egmatplotlib==3.10.8but breaks onmatplotlib-3.11.0with:ValueError: the values passed in the (value, color) pairs must increase monotonically from 0 to 1.The release notes API changes do not mention this, https://matplotlib.org/stable/api/prev_api_changes/api_changes_3.11.0.html
Code for reproduction
Actual outcome
It now fails:
Expected outcome
This used to work, eg
Additional information
No response
Operating system
No response
Matplotlib Version
3.11.0
Matplotlib Backend
macosx (or Linux tested via CI)
Python version
3.12.10 (and Python 3.11 and 3.13 via CI)
Jupyter version
No response
Installation
pip