Like many others, as soon as the popular web game "Wordle" went viral in 2021/2022, I, like many others, became obsessed with it. When I began the Master of Data Sciences - Computational Linguistics at UBC, one of the first potential applications I thought of with the skills I would be learning was an algorithm to be able to automatically solve the game's puzzles. Fast forward to now, I'm a few months from finishing the program and I've began developing this repo in order to tackle this question. Ironically, at this point I hadn't played the game in months and I was not aware there had since been developed the "Wordle Bot", that does exactly what I had set out to create, so this whole project was made completely independent of Wordle Bot, with no knowledge or intentional borrowings from it.
If you would like to see this backend in action, I've developed a frontend using Streamlit Cloud. Check it out here: https://wordlewizard.streamlit.app/
This folder contains all data files either used or generated by processing in the project.
This file contains cleaning and creation of datsets used throughout the project.
This file contains all proprietary both helper and main functions used throughout the project. Function documentation is provided within the file.
This file uses a generated csv of playthroughs of Wordle puzzles (generated from within the wordle_testing.ipynb
file), to define a machine learning pipeline that can predict the number of turns it would take for a user would to get from a given starting word to a given ending word.
This is just for testing things for now. Contents will be deleted/merged to other files once things are more finalized.
This has all been developed by me, exclusively, from scratch, so if you have any feedback or would like to suggest changes to current or future direction of the project, feel free to contact me at [email protected]