Xarray aply ufunc is giving new error

Cloud Pangeo Cloud Support
1

Hi, I’ve been successfully using the same block of code for a while when creating a potential density variable (sigma2) but today encountered an error I haven’t seen before when executing this code. I’ve included a simple reproduction that will give the same error I am encountering. I would appreciate anyone’s help in resolving this issue!

import xarray as xr
import numpy as np
from fastjmd95 import jmd95numba

salt = xr.DataArray(np.zeros((5, 5), dtype='float32'), dims=('y', 'x'))
temp = xr.zeros_like(salt)

sigma2 = jmd95numba.rho(salt.values, temp.values, 2000)

Here is the error message for reference:
UFuncTypeError: ufunc 'rho' did not contain a loop with signature matching types (<class 'numpy.dtype[float64]'>, <class 'numpy.dtype[float64]'>, <class 'numpy.dtype[float64]'>) -> <class 'numpy.dtype[float64]'>

2

For context, the docker image on Pangeo Cloud was just upgraded and package versions have changed. The error here doesn’t have to do with xarray–it is reproducible with just fastjmd95.

3

I got to the bottom of this and opened an issue in pangeo docker images:

4

I have used the admin configurator to point the hub at pangeo-notebook 2022.03.08. So this problem should be resolved.

5

Yes! It’s working again, thank you for the swift fix!