Skip to content

Mixing text and graphics in PDF

3 messages · Wright, Kevin, Peter Dalgaard, A.J. Rossini

Wright, Kevin
# 
I guess I should have emphasize the word "easy" in the original post.  I am
aware of Sweave and have used it to produce terrific quality output, but it
takes quite a bit of effort to get that quality (setting up tex document,
erasing previous graphics, running the script, tex'ing the output, etc).

HTML code with separate graphic images seems even more work.

I was hoping for a non-presentation quality way to dump both graphics and
text straight from R into pdf file using the pdf() device.  Something akin
to the way a Mathematica or Maple session looks like.

Wouldn't it be nice, for example, to fit a linear model, then send both the
residual plots and the anova table straight to the PDF file for easy storage
in a single document?  Doesn't need to need to be fancy.

The idea of a Mathematica-style transcript session is quite appealing to me
the more I think about it, but might be too-much work for the R community to
create.  Even the commercial software S-Plus lacks such reporting
capabilities.  Think--"summary" methods could produce text AND graphics--an
intriguing idea.

Kevin Wright
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Peter Dalgaard
# 
"Wright, Kevin" <kevin.wright at pioneer.com> writes:
HTML output is not all that far-fetched. Or perhaps XML. We could in
principle have all our print methods replaced by HTML output methods
(we might need to exercise a little more discipline to get rid of
explicit cat() and print() statements, but I'm not even sure that is
necessary). The default output method could simply be a plaintext
wrapper, like the HTML

<pre><kbd>
...
</kbd></pre>

and we could have some fairly obvious alignment structures for
vectors, arrays, tables, etc. It's not an afternoon's work, but it's
definitely doable. (And I wouldn't be too surprised if someone was
already playing with the idea...)
A.J. Rossini
# 
kevin> I guess I should have emphasize the word "easy" in the original post.  I am
    kevin> aware of Sweave and have used it to produce terrific quality output, but it
    kevin> takes quite a bit of effort to get that quality (setting up tex document,
    kevin> erasing previous graphics, running the script, tex'ing the output, etc).

    kevin> HTML code with separate graphic images seems even more work.

    kevin> I was hoping for a non-presentation quality way to dump both graphics and
    kevin> text straight from R into pdf file using the pdf() device.  Something akin
    kevin> to the way a Mathematica or Maple session looks like.

actually, the ESS noweb mode facilitates a great deal of this
production.  Combine it with an appropriate makefile, and document
construction (and verification of the results) is pretty trivial.

I recognize that we need to make tools like xtable cleaner to use (not
easier, just cleaner IMHO), and ESS needs to provide template
insertion for figures (and xtable runs).  That is coming.

The figures issue might be more critical -- I should probably figure
out the right incantation for the standard plot methods which produce
multiple figures...

    kevin> The idea of a Mathematica-style transcript session is quite appealing to me
    kevin> the more I think about it, but might be too-much work for the R community to
    kevin> create.  Even the commercial software S-Plus lacks such reporting
    kevin> capabilities.  Think--"summary" methods could produce text AND graphics--an
    kevin> intriguing idea.

Again, it's a fairly small step for the ESS interface; mostly template
construction and the issues raised above.  We are thinking about it!

best,
-tony