cmap

cmap

Scientific colormaps for python, with no dependencies beyond numpy.

With cmap, you can use any of the colormaps from matplotlib, cmocean, colorbrewer, crameri, seaborn, and a host of other collections in your python code, without having to install matplotlib or any other dependencies beyond numpy.

:book: See the complete catalog

There are a number of python libraries that provide or require colormaps or basic color support, but they all either depend on matplotlib, provide a specialized set of colormaps intended to extend those provided by matplotlib, or roll their own colormap solution that vendors/duplicates other libraries.

cmap is a lightweight library that provides a large collection of colormaps with no dependencies beyond numpy. It provides exports to a number of known third-party colormap objects, allowing it to be used across a wide range of python visualization libraries. The intention is to provide a library that can be used by any python library that needs colormaps, without forcing the user to install matplotlib (while still being compatible with matplotlib and other libraries that use matplotlib colormaps).

cmap is strictly typed and fully tested, with a focus on good developer experience.

Install

pip install cmap

conda install -c conda-forge cmap

Usage

See Documentation for full details.

cmap.Color

The cmap.Color object is a simple wrapper around a tuple of RGBA scalars, with a few convenience methods for converting to other color objects.

from cmap import Color

red = Color("red")  # or a variety of other "color like" inputs

cmap.Colormap

The cmap.Colormap object is a callable that can map a scalar value (or numpy array of values) to an RGBA color (or a numpy array of RGBA colors). API is intended to mimic the behavior of a matplotlib.colors.Colormap object (without requiring matplotlib)

In [1]: import cmap

# or a variety of other "colormap like" inputs
In [2]: cmap1 = cmap.Colormap(["red", "green", "blue"])

In [3]: cmap1(np.linspace(0,1,5))
Out[3]:
array([[1.        , 0.        , 0.        , 1.        ],
       [0.50393701, 0.24900417, 0.        , 1.        ],
       [0.        , 0.50196078, 0.        , 1.        ],
       [0.        , 0.24900417, 0.50393701, 1.        ],
       [0.        , 0.        , 1.        , 1.        ]])

Note that the input array must be normalized from 0-1, so if you're applying a colormap to an integer array (like an image) you must apply any contrast limits and rescale to 0-1 before passing it to a Colormap.

Third Party Library Support

The cmap.Colormap object has convenience methods that export it to a number of known third-party colormap objects, including:

• matplotlib
• napari
• vispy
• pygfx (& fastplotlib)
• plotly
• bokeh
• altair
• earthengine-api
• pyqtgraph

See documentation for details.

If you would like to see support added for a particular library, please open an issue or PR.

Alternatives

Other libraries providing colormaps:

• matplotlib
• seaborn (subclasses matplotlib)
• proplot (subclasses matplotlib)
• palettable (mostly data, import doesn't depend on matplotlib, but usage largely does)
• cmocean (mostly data, outputs matplotlib colormaps)
• colorcet (mostly data, usage requires either matplotlib or bokeh)
• cmasher (requires matplotlib)
• cmyt (requires matplotlib)
• cmcrameri (requires matplotlib, wraps https://www.fabiocrameri.ch/colourmaps/)
• distinctipy (generates distinct color sets, only requires numpy)
• Farrow & Ball Matplotlib (requires matplotlib)
• mplcyberpunk (requires matplotlib)

References and Further reading

• Choosing Colormaps in Matplotlib
• A Better Default Colormap for Matplotlib | SciPy 2015 | Nathaniel Smith and Stéfan van der Walt
• blog post for above video: https://bids.github.io/colormap/
• Origins of Colormaps, Cleve Moler, February 2, 2015
• Documenting the matplotlib colormaps, @endolith
• Color Map Advice for Scientific Visualization
• https://colorcet.com/, Peter Kovesi
• Kovesi: Good Colour Maps: How to Design Them.
• https://www.fabiocrameri.ch/colourmaps/