starcraft_predictor contains a pre-trained ML model that can be used to generate win probabilities throughout a Starcraft 2 game. The package uses .SC2Replay files to load in game metadata and generate predictions based on the state of the game at every 10 second interval.
The package is not currently on PyPi so the best way to use it locally is to clone the repo
git clone https://github.com/nedwebster/starcraft_predictor.git
The project uses uv for package management and project builds. Install the dependencies and the package using:
uv sync --reinstall
To analyse a replay, load the replay_scorer and use the score_replay() method. The load_scorer() function loads the pre-trained model, and can score replays given a path.
Run the following code in a notebook:
from starcraft_predictor import load_scorer
replay_scorer = load_scorer()
replay_scorer.score_replay("path/to/replay.SC2Replay")
The project uses streamlit to deploy the application as a simple web app. To run the streamlit app, run:
uv run streamlit run app.py
From there you can upload replays from your local machine and get win probability prediction plots.
.
├── example_data/ <-- example data to use in the tutorials
├── example_notebooks/ <-- notebook tutorials for app functionality
├── ml_analysis/ <-- adhoc analysis notebooks that guides parts of the ML project
├── scripts/ <-- python scripts used to train the model and score replays
├── app.py <-- streamlit app file for running the streamlit web app.
└── src/starcraft_predictor/ <-- main app folder/
├── modelling/ <-- code for the ml model
├── plots/ <-- code for plotting replay predictions
├── processing/ <-- code for processing data ready for the ML model
├── replays/ <-- code