oedokumaci / gale-shapley

Python implementation of the Gale-Shapley Algorithm.

Oral Ersoy Dokumacı

Quant Lead • ML Engineer • PhD Candidate

---

I am a Lead Quantitative Researcher and Developer, and Machine Learning Engineer at Vitus Commodities, and I work as a consultant at the Burning Glass Institute and the World Bank.

I am also a Ph.D. candidate in Economics at Bilkent University. My research interests are computational mechanism design and matching theory, decision theory, and applications of Machine Learning in Economics.

---

Languages and Tools

---

Usual Workflow Badges

---

Github Statistics

gale-shapley's Issues

Add type guards

This refactors the code and also helps mypy not to complain as much

See the related PEP

Add argument parsing for number of simulations in cli.py

Probably modern way to do it is via typer instead of argparse.

Pytest hook writing log

This can be potentially fixed with pytest mocks.

Config and log paths would not work in Windows

Custom deterministic input with GUI support

Add Self type for type hints

See the related PEP673.

Is it possible to use CUDA to accelerate this algorithm?

Print all preferences compact form

Printing all preferences in compact form can provide buggy output in one of the following two cases:

• If proposer or responder names are long
• If there are many proposers or responders, resulting in lines ending

Steps to reproduce the behavior:

1. Edit config.yaml to produce one of the two cases above.
2. Run python -m gale_shapley
3. See logs

Print a table like when it is not one of the two cases above.

All of the below systems can produce the bug

• Python version: 3.9, 3.10, 3.11
• OS: Ubuntu/MacOS/Windows

Installing for the first time

When installing the project for the first time, logs will produce error.

Steps to reproduce the behavior:
Setup the project following Developer Guide in README.md in a fresh environment.

All of the below systems can produce the bug

• Python version: 3.9, 3.10, 3.11
• OS: Ubuntu/MacOS/Windows