problems with R graphics

problems with R graphics

[Matthew Landis]

Hello all,

I believe this is my first post on this list, and I'm very new to org-babel, so 
forgive me if I am posting nonsense.  However, I'm really enthused about the 
possibilities org-babel presents.  Unfortunately, I can't seem to get graphics 
to insert themselves properly on export.  I've included an example that 
reproduces my problem below.  For what it's worth, I am using Windows 7 with a 
freshly downloaded issue of Vincent Goulet's package.  Session info:

Emacs version: GNU Emacs 23.4.1 (i386-mingw-nt6.1.7601) of 2012-02-04 on MARVIN
Org mode version 7.8.03
ESS version 5.14
R version 2.14.1

Many thanks for any thoughts or requests for more info.

Matt

org example
----------------------------

#+TITLE: An attempt to make org-mode work with R

In this example, I want to fit a regression and plot it.  I'm
following the examples at the [[http://orgmode.org/worg/org-
contrib/babel/languages/ob-doc-R.html][org manual]] and Erik Iverson's 
[[https://github.com/erikriverson/org-mode-R-tutorial][tutorial]].

* A stand-alone code block

First, create some data, perform a regression, and plot it.  This
works great.
#+begin_src R :results graphics :file regress.png :exports both

  x <- runif(n = 100, min = 0, max = 20)
  y <- 5*x^2 + 0.3*x + 3 + rnorm(100, sd = 200)

  f <- lm(y ~ x)

  plot(x, y, bty = 'n')
  abline(f, lty = 'dashed')

#+end_src
Success.

* In code blocks using ":session"
Now do the same thing, but across an unnamed session

#+begin_src R :session :exports code :results silent

  x <- runif(n = 100, min = 0, max = 20)
  y <- 5*x^2 + 0.3*x + 3 + rnorm(100, sd = 200)

#+end_src

Now, I would like to display the results of the regression as R output.

#+begin_src R :session :results output :exports both
  f <- lm(y ~ x)
  summary(f)
#+end_src
Again, success!  This is really going to come in handy.

#+begin_src R :session :results graphics :file regress_session.png :exports both

 plot(x, y, bty = 'n')
 abline(f, lty = 'dashed')

#+end_src
This code block doesn't work. No error, but no file is created.
Also, if I press C-c C-c in the code block, in the org file I get

#+begin_example
#+results:
[[file:regress_session.png]]
#+end_example

But no file is actually created.  What could be going wrong?

Re: problems with R graphics

[Matthew Landis]

Matthew Landis <landis <at> isciences.com> writes:

> 
> Hello all,
...

> Unfortunately, I can't seem to get graphics 
> to insert themselves properly on export.  

....

> org example
> ----------------------------
...
> * A stand-alone code block
> 
> First, create some data, perform a regression, and plot it.  This
> works great.
> #+begin_src R :results graphics :file regress.png :exports both
> 
>   x <- runif(n = 100, min = 0, max = 20)
>   y <- 5*x^2 + 0.3*x + 3 + rnorm(100, sd = 200)
> 
>   f <- lm(y ~ x)
> 
>   plot(x, y, bty = 'n')
>   abline(f, lty = 'dashed')
> 
> #+end_src
> Success.
...

> #+begin_src R :session :results graphics :file regress_session.png :exports 
both
> 
>  plot(x, y, bty = 'n')
>  abline(f, lty = 'dashed')
> 
> #+end_src
> This code block doesn't work. No error, but no file is created.

...

I actually figured this one out just now.  If creating a file in a 'stand alone' 
code block, the default directory is the directory that the org file itself is 
in.  However, in a 'session' code block, the default directory is something else 
(R's default?).  If I specify the full path to the file in the 'session' code 
block, things work better.

Sorry for the noise!

M

Re: problems with R graphics

[Nick Dokos]

Matthew Landis <landis@isciences.com> wrote:

> Matthew Landis <landis <at> isciences.com> writes:
> 
> ...
> 
> > #+begin_src R :session :results graphics :file regress_session.png :exports 
> both
> > 
> >  plot(x, y, bty = 'n')
> >  abline(f, lty = 'dashed')
> > 
> > #+end_src
> > This code block doesn't work. No error, but no file is created.
> 
> ...
> 
> I actually figured this one out just now.  If creating a file in a 'stand alone' 
> code block, the default directory is the directory that the org file itself is 
> in.  However, in a 'session' code block, the default directory is something else 
> (R's default?).  If I specify the full path to the file in the 'session' code 
> block, things work better.
> 

Maybe Windows is doing something different, but on Linux, both plots are
produced with no problem, both in the current directory.

Nick