COLOR ANALYZER

Allows to get the 3 most dominant "dummy" colors from image files in given directory. "Dummy" means one of following: black, red, yellow, green, cyan, blue, magenta, white.

This script uses fengsp's Color Thief as image color analyzing tool and Flask as a simple HTTP API server. py.test is used as a test runner.

Implementation:

Script:

Color Thief's API is used in this script by calling color_thief.get_palette on each file. Returned color palette is used as a base for presenting the result of 3 "dummy" colors most dominant in this file. Of course, two or more colors from Color Thief's palette can represent the same "dummy" color, therefore get_palette is called with argument color_count equal to 10 - from 10 top colors we try to find 3 unique "dummy" colors. #####Caution: in some cases (E.g. black and white pictures) only 2 results are expected.#####

API:

The simplest way to provide a simple API endpoint was to use Flask. One small file with one function is enough to have a valid API endpoint and that's all that was needed.

Installation:

pip install -r requirements.txt

Usage:

Script:

python analyzer.py
>>> There are no image files!

Ooops!

python analyzer.py path-to-image-directory
>>> data/magical.png:
> red
> black
> white
>>> data/all-the-sweaters.jpg:
> white
> black
> red

API:

export FLASK_APP=endpoint.py
flask run -p 8000
curl -F "file=@data/magical.png" 127.0.0.1:8000/find_colors.json
{"colors": ["red", "black", "white"]}

Test:

pip install -r requirements_test.txt
pytest test.py

Also, there are two example image files in data directory.

Possible improvements:

Images could be analyzed asynchronously using Gevent or Celery which would significantly improve runtime and stability, but that wouldn't count as "an easy script".