Skip to content

OpenSportsLab/DatasetAnnotationTool

Repository files navigation

OSL Dataset Visualizer

Documentation Status

This is a simple PyQt5-based visualizer for datasets in the OSL (Open Sports Lab) JSON format.

Features

  • Load and view OSL JSON files.
  • Display dataset contents in a user-friendly interface.
  • Easy to extend with visual overlays and timeline viewers.

🔧 Environment Setup

We recommend using Anaconda or Miniconda for managing your Python environment.

Step 1 – Create a new Conda environment

conda create -n osl-visualizer python=3.9 -y
conda activate osl-visualizer

Step 2 – Install dependencies

pip install pyqt6

Optionally, if you plan to add video rendering or computer vision:

pip install opencv-python

🚀 Run the Visualizer

From the root of the project folder, launch the app with:

python osl_visualizer/main.py

A window will open where you can load your OSL JSON files.

Tips: Download a dataset from HF and run the app with that dataset

python tools/download_osl_hf.py \
--url https://huggingface.co/datasets/OpenSportsLab/HistWC/blob/main/HistWC-finals.json \
--output-dir /Users/giancos/Documents/HistWC/
python osl_visualizer/main.py --osl_file /Users/giancos/Documents/HistWC/HistWC-finals.json

🚀 Run the Installer

This script is automatically called whenever you push to Github

pip install pyinstaller

pyinstaller --onefile --windowed main.py


Build the docs

pip install mkdocs mkdocs-material mkdocstrings[python]
mkdocs gh-deploy

📄 License

This project is open source and free to use under the MIT License.

About

GUI for manipulation of OSL Action Spotting datasets

Resources

License

Contributing

Stars

3 stars

Watchers

0 watching

Forks

Packages

 
 
 

Contributors

Languages