Recipe Roulette makes meal planning easier with machine learning and artificial intelligence.
This project is currently in development. Users can sign up, sign in, sign out, rate recipes, and receive recommendations based on their similarity to other users (collaborative filtering).
The long-term goal is to deploy a web application where users can view recommended recipes, save their own recipes, and rate recipes that they cook to fine-tune the recommendations. This may include the addition of item-based filtering or deep learning models using ingredient and taste vectors.
Recipe Roulette is a Flask application tested against Python >= 3.6.
To install Recipe Roulette, clone this repo and run:
pip install -e .
After you install the package, initialize and seed the database:
flask init-db
The script will automatically import ~20 recipes from AllRecipes.com.
Users can sign up and rate recipes based on the title, image, and link.
As ratings accumulate, recommendations will begin to appear under the recipe to rate.
You can configure multiple forms of machine learning by passing a different argument into the getRecommendations function in the rating.py blueprint.
The current options are:
- Euclidean Distance:
similarity=sim_distance - Pearson Coefficient:
similarity=sim_pearson
If you're interested in this line of work, please contact me or even submit a pull request to this project.
Recipe Roulette is licensed under the MIT License.