Hi all, I was wondering if anyone knows roughly when to expect a pip-installable version of pygplates that is compatible with Python 3.13 on MacOS?
It will happen once Macports supports the boost library for Python 3.13. I think it’s being held up by their transition to NumPy 2.
Hopefully it’s relatively soon. In the meantime you can install pygplates on conda for Python 3.13.
A note for the record: This is different than this post where we were going to wait for Macports to support NumPy 2 (so that pyGPlates would work on NumPy 1.x and 2.x). But turns out we didn’t need to wait. That’s because pyGPlates uses the NumPy C-API (not boost-python-numpy which comes from Macports). So using NumPy 1.x (in boost-python-numpy) on Macports was fine - we could still separately install NumPy 2.x as a Python package (when building pyGPlates). But pyGPlates still uses boost-python (ie, without the -numpy), and that’s only supported for Python versions <= 3.12 (on Macports). It appears that Python 3.13 requires NumPy 2.1 (hence the hold-up while Macports transitions to NumPy 2 for Python 3.13).
There are now pip-installable versions of pyGPlates 1.0 for Python 3.13 on macOS (Arm and Intel architectures). (Note: Macports recently added Python 3.13 to Boost 1.76)
So that completes the binary wheels for pyGPlates 1.0 (ie, there’s support for Python versions 3.8, 3.9, 3.10, 3.11, 3.12, 3.13 on Linux Arm, Linux Intel, macOS Arm, macOS Intel, and Windows Intel).
2 Likes