The ChessAIThon project (2025-1-ES01-KA220-VET-000354329) is co-funded by the European Union. The views and opinions expressed in this publication are those of the author(s) only and do not necessarily reflect those of the European Union or the Spanish Service for the Internationalisation of Education (SEPIE). Neither the European Union nor the National Agency SEPIE can be held responsible for them.

Chess data: formats structures and dataset files, control version

Home

Table of Contents

Part 2

Sharing Datasets Publicly

To share their chess datasets publicly, students should follow these basic steps:

  • Create a new public repository on GitHub.
  • Commit their prepared files.
  • Use the git push command to upload their data, making it available to the global open-source community.Automating the Pipeline with GitHub Actions

Explain that GitHub Actions automates the ETL (Extract, Transform, Load) process, linking human-readable data with the format AI needs.

  • The Trigger & Transformation: When a new CSV (for humans) is pushed to GitHub, the Action automatically runs the Python ETL code (using Pandas/NumPy) inside a Colab Notebook. It converts the data to the 77x8x8 numerical representation (our AlphaZero format) and indexes the moves (0 to 4096).
  • The Delivery: The Action saves the result as a compressed Parquet file (for AI training) and pushes it back to the repository. The AI training notebook then reads this optimized data directly from GitHub.

This demonstrates a professional CI/CD (Continuous Integration/Continuous Delivery) pipeline, teaching students to manage and automatically prepare data for a large-scale AI project.