We are using hyperopt-sklearn in Databricks environment, installing using pip install git+https://github.com/hyperopt/hyperopt-sklearn pulls in numpy 2.0, which in Databricks, but I guess possibly elsewhere, breaks a lot of things. This is because the setup.cfg does not contain the same upper bounds on packages as requirements.txt
I have simply updated the setup.cfg to be consistent with requirements.txt, if you think different upper bounds should be tested and used I can invest some time into it.
I'm also curious why pandas is not listed as dependency in setup.cfg?
Hello
We are using hyperopt-sklearn in Databricks environment, installing using
pip install git+https://github.com/hyperopt/hyperopt-sklearn
pulls in numpy 2.0, which in Databricks, but I guess possibly elsewhere, breaks a lot of things. This is because thesetup.cfg
does not contain the same upper bounds on packages asrequirements.txt
I have simply updated the
setup.cfg
to be consistent withrequirements.txt
, if you think different upper bounds should be tested and used I can invest some time into it.I'm also curious why pandas is not listed as dependency in
setup.cfg
?