nannernest

A small package for optimizing banana coverage on peanut butter and banana sandwiches. See blog post for more info.

Installation

Python 3.7+ is required.

nannernest is generally pip installable. Due to some C dependencies with the nesting library that I use nest2D, along with an outstanding PR, I would recommend the following way to install everything:

First, make sure you have boost and cmake installed. If you are on Linux, then you may have cmake installed, and you can install boost with

sudo apt-get install libboost-all-dev 

I'm on a Mac, and I installed cmake with conda and boost with brew:

conda install cmake
brew install boost

Next, pip install my fork of nest2D:

pip install git+https://github.com/EthanRosenthal/[email protected]

Finally, pip install nannernest

pip install nannernest

Usage

Take a top-down picture that contains your banana and at least one slice of bread. Pass the image in via command line:

$ nannernest my_image.jpg

CLI Details

$ nannernest --help
Usage: nannernest [OPTIONS] IMAGE_PATH

Arguments:
  IMAGE_PATH  Image file which contains bread and banana  [required]

Options:
  --num-slices INTEGER            Total number of slices to cut the banana
                                  into. This number defines the slice
                                  thickness.  [default: 22]

  --mask-threshold FLOAT          Threshold of segmentation mask.  [default:
                                  0.6]

  --peel-scaler FLOAT             Fraction of slice that is assumed to belong
                                  to banana insides versus the peel.
                                  [default: 0.8]

  --ellipse-ratio FLOAT           Assumed ratio of minor axis to major axis of
                                  banana slice ellipses  [default: 0.85]

  --plot-segmentation / --no-plot-segmentation
                                  Whether or not to plot the segmentation
                                  masks  [default: False]

  --plot-slicing / --no-plot-slicing
                                  Whether or not to plot the slicing circle
                                  and skeleton  [default: False]

  --output TEXT                   Name of file to output  [default:
                                  perfect_sandwich.jpg]