======= Gradify

A python library to generate CSS gradient from an image. This is a fork of https://github.com/fraser-hemp/gradify. The original gradify project had only a CLI interface. With this fork, you can use gradify as a library in your Python code.

.. image:: https://user-images.githubusercontent.com/2115303/35187613-c6fe6fe8-fe3b-11e7-9b9d-3e088e460a1d.jpg

Installation

Install the latest release from PyPI:

.. code-block:: sh

pip install gradify

Usage

.. code-block:: python

gradify.generate_css(fp, single_color=False, use_color_spread=False)

• fp: a filename (string), pathlib.Path object or a file object. The file object must implement read(), seek(), and tell() methods, and be opened in binary mode.

• single_color: only produce a single, uniform background color - this is much quicker and has all browser support

• use_color_spread: this flag will give the color which has the least spread over the image the highest priority when assigning directions (opposed to most dominant color). This feature improves overall accuracy, however adds complexity and in unique cases it produces counter-intuitive results

Example:

.. code-block:: python

import gradify
css = gradify.generate_css('sample.png')

Advanced Usage

You can use Gradify class for advanced usage.

.. code-block:: python

from gradify import Gradify

# simple usage
g = Gradify('sample.png', single_color=False, use_color_spread=False)
css = g.generate_css()

# advanced usage
g = Gradify('sample.png', single_color=False, use_color_spread=False, black_sensitivity=4.3,
            white_sensitivity=3, num_colors=4, resize=55, uniformness=7, use_prefixes=False)
css = g.generate_css()

Default parameters produce good result.

From original repo:

The only suggestion is increasing the uniformness (by lowering it's value). It can improve
the general case, improve speed, but decrease the upper limits of accuracy.
Increasing sensitivity to black will do the same.

Credits

Many thanks to Fraser Hemphill_ for writing the original gradify project.

License

MIT

.. _Fraser Hemphill: https://github.com/fraser-hemp