Closed yanwengong closed 3 years ago
Hi,
Can you try to install the Giotto environment like installGiottoEnvironment(force_miniconda = TRUE)
from a fresh R session? So without loading any other library (besided Giotto) or running a prior analysis. This should normally work.
If that works I can further try to figure out why you can not provide your own python path. That should normally work. Which reticulate version are you using?
Hi. I tried installGiottoEnvironment(force_miniconda = TRUE) from a fresh R session, but still get " Error: 'miniconda_path' is not an exported object from 'namespace:reticulate'"
packageVersion("reticulate") [1] ‘1.13’
Ok, thanks for the feedback. I think it's related to the version of reticulate you're using. If you upgrade to the newest version it should work. I think anything > 1.15 should actually work. If this solves the issue, then I'll also update this in Giotto.
Hi. I updated reticulate to ‘1.18’ and it worked. Thanks for the suggestion!
Hi.
I created a Conda env called giotto where I install all the required python package through conda install (note I have anaconda3 instead of miniconda ).
python_path <- "/Users/your_username/anaconda3/envs/giotto/bin/python" instrs = createGiottoInstructions(show_plot = TRUE, save_plot = TRUE, save_dir = results_folder, python_path = python_path)
I got the error of "Error: 'miniconda_path' is not an exported object from 'namespace:reticulate'"
Then, I tried to use the following code to install the python packages automatically: python_path = NULL if(is.null(python_path)) { installGiottoEnvironment(force_miniconda = TRUE) }
Got the same error " Error: 'miniconda_path' is not an exported object from 'namespace:reticulate' "
Do you know what's happening?