Closed jluethi closed 1 month ago
There are multiple instances of np.float32 / np.float64 being used in the code. But also twice just np.float: https://github.com/fmi-basel/gliberal-scMultipleX/blob/83e94dfa04f187969f7a0ee024220123f967669b/src/scmultiplex/platymatch/platymatch/estimate_transform/apply_transform.py#L28
Support for np.float was dropped in numpy==1.24.0
(I provisionally changed the numpy verison in the 0.5.0 Fractal task on the FMI setup to 1.23.5 to avoid this deprecation until we've fixed it)
Some of the measurement code still uses the deprecated np.float. This now leads to the following error in minor_major axis ratio measurements:
We should check where np.float is used and test whether everything works as expected with just float or whether the np.float64 is necessary.