Clone this repository into a fresh Python 3.13 environment:
git clone https://github.com/PJ-Watson/niriss-tools.git
Install using:
python -m pip install ./niriss-tools
Navigate to line 177 of site-packages/stsci/tools/stpyfits.py, and change the dtype to int:
raw_data = np.zeros(shape=dims, dtype=int) + pixval
Install adelie as an alternative NNLS solver. This can provide a significant speed increase over scipy for higher-dimensional problems. Installation instructions can be found here.
You can, but I can't guarantee the package dependencies will resolve properly. If you manage to make it work, let me know and I'll update the installation instructions.
Probably. Although you may find that some existing packages have been reinstalled (almost certainly grizli, drizzlepac, and photutils).
In Numpy 2.0, type promotion in this manner throws an exception. Until I (or someone else) checks that it isn't grizli or drizzlepac at fault here, this is the quickest solution.