marcotinacci / wordle

Python implementation of different types of Wordle strategies

Wordle

Python implementation of a Wordle player. The source code provides a structure to easily create Wordle players, and some built-in strategies:

• heuristic: a simple heuristic strategy that priorities words that contain letters with the highest frequency. Linear time complexity $O(W)$.
• minmax: minmax strategy that minimize the maximum number of candidates on the next move. Quadratic time complexity $O(W^2)$.
• precomputed: decorator strategy that precomputes the full decision tree. Constant time complexity $O(1)$ at runtime. Precomputation complexity is the same as the targeted strategy, as it is repeated for each step of the decision tree which has a constant number of nodes.

Installation

Clone the repository and run the following command from the project root to install the required dependencies (use a virtual environment if you wish, or don't, I'm not your mother):

pip install -r requirments.txt

Usage

The package provides a CLI interface to play with the different functions. Check the help for more information in this way:

$ python cli.py --help

The main subcommand is play, that let you play with the different strategies:

$ python cli.py play

Here an example of a game with the minmax strategy:

$ python cli.py play -S minmax -p
hint: arise
$ python cli.py play -S minmax -p arise .____
hint: candy
$ python cli.py play -S minmax -p arise .____ candy _X__.
hint: bayou

-S let you choose the strategy to use. The default strategy is heuristic.

-p grabs the strategy from the precomputed tree named data/strategies/<strategy-name>.json.

Finally you need to provide the sequence of guess/feedback pairs, where the feedback is encoded as a string of the following characters:

• _: incorrect letter
• .: correct letter, wrong position
• X: correct letter, correct position

Acknowledgments

Thanks to @lostella for the always constructive conversations (that I'm missing), and to @davideboschetto and @albertoguiggiani for the laughs.