R help - Jul 2024 - reticulate + virtual environments

Hi,

I am using reticulate and a virtual environment (not conda) to run 
Python scripts from RStudio. However, when I try to use my own 
(existing) virtual environment, reticulate does not use it. If I run my 
scripts, the installed modules (e.g., py_install("pandas", 
"mmstat4.hu.data")) are not found. I believe this happens because 
reticulate is using r-reticulate instead of mmstat4.hu.data. How can I 
force reticulate to use my virtual environment?

Thanks Sigbert

 > library("reticulate")

 > py_config()

python:         /home/sk/.virtualenvs/r-reticulate/bin/python

libpython: 
/usr/lib/python3.10/config-3.10-x86_64-linux-gnu/libpython3.10.so

pythonhome: 
/home/sk/.virtualenvs/r-reticulate:/home/sk/.virtualenvs/r-reticulate

version:        3.10.12 (main, Mar 22 2024, 16:50:05) [GCC 11.4.0]

numpy: 
/home/sk/.virtualenvs/r-reticulate/lib/python3.10/site-packages/numpy

numpy_version:  2.0.0

 > use_virtualenv("mmstat4.hu.data", required = TRUE)

 > py_config()

python:         /home/sk/.virtualenvs/r-reticulate/bin/python

libpython: 
/usr/lib/python3.10/config-3.10-x86_64-linux-gnu/libpython3.10.so

pythonhome: 
/home/sk/.virtualenvs/r-reticulate:/home/sk/.virtualenvs/r-reticulate

version:        3.10.12 (main, Mar 22 2024, 16:50:05) [GCC 11.4.0]

numpy: 
/home/sk/.virtualenvs/r-reticulate/lib/python3.10/site-packages/numpy

numpy_version:  2.0.0

-- 
https://hu.berlin/sk
https://hu.berlin/mmstat

Have you tried https://rstudio.github.io/reticulate/  ?

Generally speaking, complex nonstandard package specific questions
such as yours rarely get a reply here -- there are 20,000+ packages
(and counting) after all! As reticulate was created by and integrated
with RStudio/Posit, I would think their site and help resources might
be a better venue. Of course, if you don't use RStudio, you may have
no joy there either.

Cheers,
Bert




On Sun, Jul 14, 2024 at 10:56?PM Sigbert Klinke
<sigbert at wiwi.hu-berlin.de> wrote:
>
> Hi,
>
> I am using reticulate and a virtual environment (not conda) to run
> Python scripts from RStudio. However, when I try to use my own
> (existing) virtual environment, reticulate does not use it. If I run my
> scripts, the installed modules (e.g., py_install("pandas",
> "mmstat4.hu.data")) are not found. I believe this happens because
> reticulate is using r-reticulate instead of mmstat4.hu.data. How can I
> force reticulate to use my virtual environment?
>
> Thanks Sigbert
>
>  > library("reticulate")
>
>  > py_config()
>
> python:         /home/sk/.virtualenvs/r-reticulate/bin/python
>
> libpython:
> /usr/lib/python3.10/config-3.10-x86_64-linux-gnu/libpython3.10.so
>
> pythonhome:
> /home/sk/.virtualenvs/r-reticulate:/home/sk/.virtualenvs/r-reticulate
>
> version:        3.10.12 (main, Mar 22 2024, 16:50:05) [GCC 11.4.0]
>
> numpy:
> /home/sk/.virtualenvs/r-reticulate/lib/python3.10/site-packages/numpy
>
> numpy_version:  2.0.0
>
>  > use_virtualenv("mmstat4.hu.data", required = TRUE)
>
>  > py_config()
>
> python:         /home/sk/.virtualenvs/r-reticulate/bin/python
>
> libpython:
> /usr/lib/python3.10/config-3.10-x86_64-linux-gnu/libpython3.10.so
>
> pythonhome:
> /home/sk/.virtualenvs/r-reticulate:/home/sk/.virtualenvs/r-reticulate
>
> version:        3.10.12 (main, Mar 22 2024, 16:50:05) [GCC 11.4.0]
>
> numpy:
> /home/sk/.virtualenvs/r-reticulate/lib/python3.10/site-packages/numpy
>
> numpy_version:  2.0.0
>
> --
> https://hu.berlin/sk
> https://hu.berlin/mmstat
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

? Mon, 15 Jul 2024 07:56:17 +0200
Sigbert Klinke <sigbert at wiwi.hu-berlin.de> ?????:
>  > py_config()  
>  > use_virtualenv("mmstat4.hu.data", required = TRUE)  
Does it help _not_ to call py_config() before use_virtualenv()?

help(py_config) says that it forces the initialization of Python. When
you later try to ask for a different virtual environment, no conflict
is detected because
normalizePath('/home/sk/.virtualenvs/r-reticulate/bin/python') is
identical to
normalizePath('/home/sk/.virtualenvs/mmstat4.hu.data/bin/python'): they
must be both symlinks to /usr/bin/python3, so reticulate is likely
thinking that it's the same Python. Thus Python is not initialised
again, but you also don't see an error.

-- 
Best regards,
Ivan