Wordle Env: A Daily Word Environment for Reinforcement Learning

Setup

Steps:

1. git pull [email protected]:alex-nooj/wordle_env.git
2. From the wordle_env directory:
   • pip install -e .
   • pip install -r requirements.txt

What's Included:

WordleCharEnv

A Wordle Gym Environment that takes actions one character at a time.

The environment will provide a mask that lets the player/agent know which letters it can guess next. For example, if the agent guesses "q", then the mask will return that the only valid letters to guess next will be specific vowels, such as "a" and "u". Once the agent guesses "u", then the mask will include all the letters that can form a word that starts with "qu", and so on.

The environment will also make sure that a letter that leads to an invalid letter cannot be entered. In the case that this does occur, the environment will remain on the current state.

WordleWordEnv

A Wordle Gym Environment that takes actions a whole word at a time.

Similar to the WordleCharEnv, if an invalid word is entered into the WordleWordEnv, the environment will produce a warning and remain on the current state.