Assuming PR #49 is merged, the only pinning we're doing in pyproject.toml is numpy<2. A PR that reverses the pinning (numpy>=2) along with the test data regenerated for the new pinning should be the way forward I think. (numpy 1.x lasted 15 years, so there's some long-term value in pinning to 2 and above).
Assuming PR #49 is merged, the only pinning we're doing in
pyproject.tomlisnumpy<2. A PR that reverses the pinning (numpy>=2) along with the test data regenerated for the new pinning should be the way forward I think. (numpy 1.x lasted 15 years, so there's some long-term value in pinning to 2 and above).