Hello,
We rely heavily on comet-maths in our HyperInSpace Community Processor (HyperCP). I recently updated our environment for the first time in a while and having been chasing down major and minor discrepancies associated with the updated packages. A presumably minor issue is now frequently raised with our update to numpy 2.4.3 in which a user warning is thrown for your use of "where" in matrix_converstion.py:
/Users/daurin/miniforge3/envs/hypercp/lib/python3.14/site-packages/comet_maths/linear_algebra/matrix_conversion.py:32: UserWarning: 'where' used without 'out', expect unitialized memory in output. If this is intentional, use out=None.
correlation = np.divide(covariance, outer_v, where=outer_v != 0)
Thanks for the great package!
Dirk
NASA/GSFC, MSU
Hello,
We rely heavily on comet-maths in our HyperInSpace Community Processor (HyperCP). I recently updated our environment for the first time in a while and having been chasing down major and minor discrepancies associated with the updated packages. A presumably minor issue is now frequently raised with our update to numpy 2.4.3 in which a user warning is thrown for your use of "where" in matrix_converstion.py:
Thanks for the great package!
Dirk
NASA/GSFC, MSU