Org babel R command session

Org babel R command session

[Sam]

I just noticed an inconsistency between the way ob-R and ob-python
select executables. For python, there are 3 variables:
org-babel-python-command, org-babel-python-command-session and
org-babel-python-command-nonsession, with the former overriding the
latter two.
 
For R, there is only org-babel-R-command, which is ignored when running
as a session.
 
Is it worth trying to make the R functionality as similar as possible to
the python? If so then I'm happy to submit a patch doing just
that. However, I'm not sure what is the best way to handle giving
command-line arguments to the R interpreter when initialising a session,
because you need to set inferior-ess-program to the executable
specifically, and put any arguments in inferior-R-args.
 
Cheers,
Sam

Re: Org babel R command session

[Ihor Radchenko]

Sam <sam@crawf.uk> writes:

> I just noticed an inconsistency between the way ob-R and ob-python
> select executables. For python, there are 3 variables:
> org-babel-python-command, org-babel-python-command-session and
> org-babel-python-command-nonsession, with the former overriding the
> latter two.

This was added because it is relatively common to use different commands
for session and non-session python blocks (things like ipython
vs. python).
  
> For R, there is only org-babel-R-command, which is ignored when running
> as a session.
> 
> Is it worth trying to make the R functionality as similar as possible to
> the python? If so then I'm happy to submit a patch doing just
> that. However, I'm not sure what is the best way to handle giving
> command-line arguments to the R interpreter when initialising a session,
> because you need to set inferior-ess-program to the executable
> specifically, and put any arguments in inferior-R-args.

The question is whether it is practically useful to have separate
commands for R session and R non-session blocks. For most interpreters,
it is not very useful - they use the same command.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: Org babel R command session

[Sam Crawford]

On 02/07/2024 15:35, Ihor Radchenko wrote:
> Sam <sam@crawf.uk> writes:
> 
>> I just noticed an inconsistency between the way ob-R and ob-python
>> select executables. For python, there are 3 variables:
>> org-babel-python-command, org-babel-python-command-session and
>> org-babel-python-command-nonsession, with the former overriding the
>> latter two.
> 
> This was added because it is relatively common to use different commands
> for session and non-session python blocks (things like ipython
> vs. python).
>    
>> For R, there is only org-babel-R-command, which is ignored when running
>> as a session.
>>
>> Is it worth trying to make the R functionality as similar as possible to
>> the python? If so then I'm happy to submit a patch doing just
>> that. However, I'm not sure what is the best way to handle giving
>> command-line arguments to the R interpreter when initialising a session,
>> because you need to set inferior-ess-program to the executable
>> specifically, and put any arguments in inferior-R-args.
> 
> The question is whether it is practically useful to have separate
> commands for R session and R non-session blocks. For most interpreters,
> it is not very useful - they use the same command.
> 

I agree that it's unlikely that users will want separate commands
(though tools like radian do exist).

Regardless, there is still the issue that, if :session is not none, then
babel will ignore org-babel-R-command, as it just makes a call to
run-ess-R. So even if you don't want to add more variables (which
ultimately I'd agree with), it might still be worth wrapping that call
in something like

(let ((inferior-ess-R-program (or org-babel-R-command
                                   inferior-ess-R-program)))
      ...)

Re: Org babel R command session

[Ihor Radchenko]

Sam Crawford <sam@crawf.uk> writes:

>> The question is whether it is practically useful to have separate
>> commands for R session and R non-session blocks. For most interpreters,
>> it is not very useful - they use the same command.
>> 
>
> I agree that it's unlikely that users will want separate commands
> (though tools like radian do exist).
>
> Regardless, there is still the issue that, if :session is not none, then
> babel will ignore org-babel-R-command, as it just makes a call to
> run-ess-R. So even if you don't want to add more variables (which
> ultimately I'd agree with), it might still be worth wrapping that call
> in something like
>
> (let ((inferior-ess-R-program (or org-babel-R-command
>                                    inferior-ess-R-program)))
>       ...)

Or we can just document that ESS defaults are used for sessions and ask
users to customize inferior-ess-R-program if they need something
non-standard for sessions.

The problem with `inferior-ess-R-program' is that it must be executable
name, and cannot be command with arguments, as in our default value of
org-babel-R-command: "R --slave --no-save". In the absense of people who
are actually making use of the proposed feature, I see no reasons to
complicate the code.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>