A hacker-friendly tool to create optimal soylent recipes from an ingredient set using a constraint solver.
First you should probably create a Python 2.7 virtualenv for this project and clone/fork it.
Next install the required Python packages by running
$ pip install -r requirements.txt
from the project directory.
Additionally install ortools by running
$ easy_install ortools
Then you can either set up your own DRI (Dietary Reference Intake) or use one of the one's provided in dri.py.
If you wish to set up your own, just copy/modify an already existing one from dri.py.
A nutrient with just a single number is assumed to have no daily intake limit, while a tuple represents the (min, max) daily intake values.
Once you have a DRI, you can add some ingredients to ingredients.py to geneate your recipe from.
All ingredients in ingredients.py will be used to generate the recipe.
Ingredients can be added as dicts to the ingredients list.
Any nutrient that is not specified for an ingredient is assumed to be 0.
It is also important to give your ingredients unique names or the solver will likely break.
The serving size is represented in grams, and all other nutrients are listed in the units specified on http://diy.soylent.me.
If you need to know the key to use for a nutrient, all possible nutrients are listed in dri.py.
Once you have your list of ingredients, you can run the solver to generate a recipe. You must provide a DRI as follows:
$ python soylentsolver.py --dri usg_m_2700
You can see a full list of available flags with:
$ python soylentsolver.py --help
- Make the solver actually work.
or-toolsmay be a failure... - Rank recipes by closeness to DRI, maybe multiple iterations to find most optimal recipes.
- Flag to minimize number of ingredients
- Potentially integration with http://diy.soylent.me so that recipes and DRI's can be automatically pushed/pulled.