================
colorobject
================
---------------
Highlights
---------------
*colorobjects* are *r, g, b, a* tuples that support easy conversion to other
color types (e.g. HSV, HLS, hexadecimal, or Matplotlib color names) and convenient
transformation (e.g. adjust color lightness, hue, saturation, or alpha).
Conversions are as convenient as *c = somecolor.hsv*, and transformations as easy as
*somecolor.lightness = c*. And because a *colorobject* is just an extension of
the list class, they can be passed directly to any Matplotlib plotting command.
*colorlists* are lists of *colorobjects* that allow you to easily modify or
construct colormaps (see module *colormaps*) or cycle through lists of maximally
distinct colors (see module *discretecolors*). Changing the brightness of a
Matplotlib colormap can be as simple as *cmap_adjust(mycolormap, 'lighten', 0.5)*
and you can even construct a colormap directly from any valid Matplotlib color
name with::
colorlist(color('Aqua', 'name'), color('Gold', 'name')).LinearSegmentedColormap()
Some examples:
* off-the-shelf Matplotlib colormap ('BuPu')
* *cmap_lighten()*
* *cmap_darken()*
* *cmap_adjust()* used to rotate the hue
* *cmap_adjust()* used to change the saturation
* constructing an entirely new colormap using the Matplotlib color names 'Aqua'
and 'Gold' as endpoints
.. image:: https://dl.dropboxusercontent.com/u/35392962/colormap_demonstration.jpg
:width: 683
:height: 202
Looking for a set of easily distinguishable colors? You could rely on one of the
lists contained in *custom_colorsets*. Access to the color sets is through a
statement such as::
colorobject.custom_colorsets.cartercarter_6
The included color sets are:
* *cartercarter_6*: 6 high contrast colors from Carter and Carter (1982)
* *greenarmytage_26*: 26 colors intended for use against a white background from
Green-Armytage (2010)
* *tatarize_269*: 269 colors from the author of the `godsnotwheregodsnot
<http://godsnotwheregodsnot.blogspot.ru/2013/11/kmeans-color-quantization-seeding.html>`_
blog
.. image:: https://dl.dropboxusercontent.com/u/35392962/discretecolors_demonstration.jpg
:width: 683
:height: 100
----------------
Installation
----------------
*colorobject* has only been tested in Python 2.7
Install through pip::
$ pip install colorobject
Requires the following non-standard libraries:
* *matplotlib*
Because the preferred installation of matplotlib can vary depending on the operating system,
matplotlib will not automatically be installed as a dependency. Instead, installation will
raise an exception if matplotlib cannot be found in the pythonpath. In this case, please
install matplotlib via your preferred method, most of which are explained `by matplotlib
<http://matplotlib.org/users/installing.html>`_
-------------------
Bonus!
-------------------
Also included are a set of custom Matplotlib colormaps. These colormaps were designed to
have every color in the colormap be clearly distinguishable against a white background. Also,
the ordering of the colors from low to high is intended to be as intuitive as possible and
there are no sudden peaks in the lightness or hue. Access to the colormaps is through a
statement such as::
colorobject.custom_colormaps.sparse2dense
The intended use for these colormaps is:
* cool2hot: temperature or heat
* deep2shallow: bathymetry maps
* dirty2clean: environmental condition or degredation
* sparse2dense: density maps
* low2high: generic
.. image:: https://dl.dropboxusercontent.com/u/35392962/custom_colormaps.jpg
:width: 683
:height: 168
--------------
Changelog
--------------
**0.8.1 (DECEMBER/02/2014)**
* Moved custom color sets from module *discretecolors* to named tuple *custom_colorsets* in
module *data*. Access is through a statement such as
*colorobject.custom_colorsets.cartercarter_6*
* Added custom color maps with named tuple *custom_colormaps* to module *data*. Access is
through a statement such as *colorobject.custom_colormaps.sparse2dense*
**0.8.0 (DECEMBER/01/2014)**
* First release