Dear Øystein, Thank you for the positive feedback! I'll look into adding support for cube decision evaluations — I’ve created an issue to track this here: 🔗 Add Cube Decision Evaluation Support · Issue #5<https://github.com/reayd-falmouth/gnubg-nn-pypi/issues/5> Regarding attribution, I’ve credited Joseph Heled in the acknowledgements section of the README<https://github.com/reayd-falmouth/gnubg-nn-pypi/blob/1a86ebb86869aa5782086096a65c4bfe31c17945/README.md?plain=1#L78> and opened a corresponding issue to track that as well: 🔗 Credit Joseph Heled as Original Author · Issue #6<https://github.com/reayd-falmouth/gnubg-nn-pypi/issues/6> And have contacted the university to ask for their guidance on licensing and attribution, as they may be able to offer further advice. Let me know if there’s anything else I should address. Best regards, David ________________________________ From: Øystein Schønning-Johansen <oyste...@gmail.com> Sent: Tuesday, May 20, 2025 11:21 AM To: DAVID REAY <dr323...@falmouth.ac.uk> Cc: bug-gnubg <bug-gnubg@gnu.org> Subject: Re: Initial Release of gnubg on TestPyPI
CAUTION: This email originated from outside of the organisation. Do not click links or open attachments unless you recognise the sender and know the content is safe. I like it! The only thing I do miss is the possibility to evaluate a cube decision. Is there a method for that? Be aware that the code you have wrapped here, is more or less Joseph Heled only code. He developed this tool to rollout the benchmark and the training databases. He has the copyright of the original code (even though it is not stated explicitly). I guess we can assume that Joseph wanted the code to be licensed under GPL... (anyone have any thoughts?). and I therefore suggest that you credit Joseph specifically and then release this wrapper under GPL. (OK! I see that it's already GPLv3) -Øystein tir. 20. mai 2025 kl. 08:50 skrev DAVID REAY <dr323...@falmouth.ac.uk<mailto:dr323...@falmouth.ac.uk>>: Dear all, I’ve completed the initial implementation of the gnubg Python package and published it to TestPyPI: 🔗 https://test.pypi.org/project/gnubg/ This builds upon the existing pygnubg program by Joseph Heled wraps the GNUBG neural network evaluation engine and exposes it to Python3 as a module installed via pip. The project is now in a reviewable state, and I’d welcome your feedback before progressing to a full PyPI release. What's Included * ⚙️ Prebuilt binary wheels for Python 3 on: * Linux (manylinux) * macOS (Intel and ARM) * Windows * 🐍 Pythonic API using snake_case naming conventions * 🔬 Core evaluation and board utilities ported from the GNUBG C++ codebase * ❗ training.cc from the original py/ directory is not yet ported * 🧪 Basic test suite included * 🔄 CI/CD workflows for building and testing (GitHub Actions) * 🚨 Linting is still failing – currently being addressed * 📚 API documentation is in progress https://gnubg.readthedocs.io/en/latest/ AI Code Generation A significant portion of the code, documentation, and DevOps work was generated with the assistance of generative AI tools, including: * GPT-4o and o4-mini-high via OpenAI ChatGPT All outputs were reviewed and curated, but please be aware that some mistakes may still exist. ________________________________ Please take a look and let me know if there’s anything you’d like changed or added. I’ll be happy to adjust the structure, naming, or packaging details based on your feedback. Thanks again, David Reay
