Installation of python module nevergrad with R package reticulate fails

[Janpie]

I'm trying to run Robyn in Databricks but it fails since the python module nevergrad fails to install properly using the r package reticulate. This code still worked properly about 1,5 months ago so maybe the issue is on some new update in databricks?

Running on 13.3 LTS and tested now also on 14.3 clusters with R libraries Robyn and reticulate preinstalled

library(reticulate)

py_install("nevergrad")

reticulate::py_config()

py_module_available("nevergrad")

#should show TRUE now shows FALSE

py_module_available("numpy")

#shows TRUE as should

This causes the Robyn pipeline to fail. Any ideas how to fix the installation of nevergrad?

[Kaniz_Fatma]

Hi @Janpie ,  try updating `pip` with `system("pip3 install --upgrade pip")`, clearing the reticulate state using `reticulate:::rm_all_reticulate_state()`, and then install `nevergrad` directly via `pip` with `system("pip3 install nevergrad")`. Ensure that `reticulate` is using the correct Python environment by specifying the path with `reticulate::use_python("/databricks/python3/bin/python3")` and check the installation with `py_module_available("nevergrad")`. 

Let me know if you need further assistance or if there’s anything else I can help with!