Inspiration
We love Costco. Buy groceries once, eat for a lifetime. Or more realistically, the expiration date on that 50 pound bag of broccoli we swore we would finish. Sadly, such a commitment eludes even the most dedicated of us as we cook whatever fancies our mind. Enter Recipr - where our big data platform (hosted by Linode!) draws from a source of over 1.5 million recipes to make sure that those greens don't go to waste.
How it works
Simply enter what you have, and how much of each item has been sitting around in your fridge, and we'll plan out your meals for you!
Challenges I ran into
Recipr has the ability to not only tell you what "lemon, sugar, and water" could make, but what and how many recipes you could make out of "15 lemons, 20 cups of sugar, and 50 cups of water" - something unmatched by over fifteen recipe-finder websites due to the problem being an NP-Complete optimization problem! Creating an algorithm that could compute this optimization a web-friendly time was a difficult obstacle the Recipr team faced. Further, designing a system to efficiently gather recipes from a database of over 1.5 million recipes was quite the worthy adversary
Accomplishments that I'm proud of
Last year at LA Hacks, we managed to discover how to destroy the Pebble smartwatch via memory-leaks. Thus, with this project, we're most proud of doing something useful. On top of this, we solved a computationally difficult problem using a clever maximum-ratio relational strategy. This computational difficulty is masked behind our beautiful frontend and lightning-fast database transactions.
What I learned
We learned that Mongo is quick to setup, Flask is easy to use, and both Python and Javascript are really hard to debug. On top of this, we learned more about the wide variety of recipes available to the world and the beauty of services that provide the knowledge to us in a reliable manner. Mainly, as with all hackathons, we learned that working late hours is by far not the optimal strategy.
What's next for Recipr
Currently, the website thrives by allowing for quick interaction of cached resources; however, when things need to be cached, things get terribly slow. The next step for Recipr is an optimization-overhaul, where the systems are sped up, code is cleaned, and all hackathon debug statements (ie 'poop1', 'poop2') are removed. Once this is done, we plan on adding even more detail to the recipes available to the end user and enhancing their experience.
Log in or sign up for Devpost to join the conversation.