[R] lattice grob

Thu Mar 25 20:54:46 CET 2010

Hi,

On 24 March 2010 23:22, Paul Murrell <p.murrell at auckland.ac.nz> wrote:
> Hi
>
> baptiste auguie wrote:
>>
>> Thanks Felix and Paul. I had overlooked grid.grabExpr, assuming that
>> one had to draw on a device before grabbing the output.
>>
>> Now I'm not sure if there's any difference between either solution,
>> I'll go for the shortest.
>
> As Felix pointed out, one possible problem with the grid.gradExpr() approach
> is that you get a copy of what was drawn on the current device, which may be
> dependent on things like how big the current device is, what fonts it uses,
> etc.

I understand the point you're making, except for one detail:
grid.grabExpr() can capture the output without any real device being
open. I am guessing that there is some kind of virtual "NullDevice"
playing its role, but then the captured output wouldn't know to decide
anything about sizes, fonts, etc., would it?

This may well be explained in your 'R graphics' book, unfortunately I
left my copy abroad. I don't remember seeing mention of the
drawDetails() approach for this purpose though, which made me a little
wary.

My current code seems to work fine with grid.grabExpr(),
http://code.google.com/p/gridextra/source/browse/trunk/R/arrange.r

I wonder what's the reason behind most high-level packages based on
Grid not defining their output as a grob/gTree but pushing/popping
viewports directly instead.

Thanks again,

baptiste

>
> The drawDetails() approach means that 'lattice' will calculate what to draw
> every time you want to draw, so it should adapt to different device sizes
> and different device properties more gracefully.
>
> On the other hand, the drawDetails() approach only records a VERY high-level
> description of what you are drawing (a 'lattice' object), so you cannot
> fiddle about with the low-level details of what you draw. For example,
>  following  ...
>
> latticeGrob <- function(p, ...){
>  grob(p=p, ..., cl="lattice")
> }
> drawDetails.lattice <- function(x, recording=FALSE){
>  lattice:::plot.trellis(x$p, newpage=FALSE)
> }
> p1 <- xyplot(1:10 ~ 1:10)
> g1 <- latticeGrob(p1)
> grid.draw(g1)
>
> ... grid.ls() gives you ...
>
>> grid.ls()
> GRID.lattice.53
>
> ... whereas the grid.gradExpr() approach records all of the bits and pieces
> of the drawing, e.g., following ...
>
> p1 <- xyplot(1:10 ~ 1:10)
> g1 <- grid.grabExpr(print(p1))
> grid.draw(g1)
>
> ... grid.ls() gives you ...
>
>> grid.ls()
> GRID.gTree.94
>  GRID.rect.85
>  plot1.xlab
>  plot1.ylab
>  GRID.segments.86
>  GRID.segments.87
>  GRID.text.88
>  GRID.segments.89
>  GRID.text.90
>  GRID.segments.91
>  GRID.points.92
>  GRID.rect.93
>
>
> So it's a bit of a trade-off.
>
> Paul
>
>
>> Best,
>>
>> baptiste
>>
>> On 22 March 2010 00:18, Felix Andrews <felix at nfrac.org> wrote:
>>>
>>> What's wrong with using grid.grabExpr?
>>>
>>> p1 <- xyplot(1:10 ~ 1:10)
>>> g1 <- grid.grabExpr(print(p1))
>>>
>>> I can imagine there would be potential problems to do with the
>>> plot-time aspect and layout calculations...
>>>
>>>
>>>
>>> On 19 March 2010 21:51, baptiste auguie <baptiste.auguie at googlemail.com>
>>> wrote:
>>>>
>>>> Dear list,
>>>>
>>>> I'm trying to arrange various grid objects on a page using a
>>>> frameGrob. It works fine with basic grobs (textGrob, gTree, etc.), and
>>>> also with ggplot2 objects using the ggplotGrob() function. I am
>>>> however stuck with lattice. As far as I understand, lattice produces a
>>>> list of class trellis, which is eventually displayed using the
>>>> plot.trellis method. I am not sure if/how one can convert this list
>>>> into a high-level grob. I tried the following,
>>>>
>>>> latticeGrob <- function(p, ...){
>>>>  grob(p=p, ..., cl="lattice")
>>>> }
>>>>
>>>> drawDetails.lattice <- function(x, recording=FALSE){
>>>>  lattice:::plot.trellis(x$p)
>>>> }
>>>>
>>>> p1 <- xyplot(1:10 ~ 1:10)
>>>> g1 <- latticeGrob(p1)
>>>>
>>>> grid.draw(g1) # works fine
>>>>
>>>> but,
>>>>
>>>> fg <- frameGrob(layout = grid.layout(1,1))
>>>> fg <- placeGrob(fg, g1, row = 1, col = 1)
>>>> grid.draw(fg)
>>>>
>>>> Error in UseMethod("depth") :
>>>>  no applicable method for 'depth' applied to an object of class "NULL"
>>>>
>>>> Ideas are most welcome,
>>>>
>>>> Best regards,
>>>>
>>>> baptiste
>>>>
>>>>> sessionInfo()
>>>>
>>>> R version 2.10.1 RC (2009-12-06 r50690)
>>>> i386-apple-darwin9.8.0
>>>>
>>>> locale:
>>>> [1] en_GB.UTF-8/en_GB.UTF-8/C/C/en_GB.UTF-8/en_GB.UTF-8
>>>>
>>>> attached base packages:
>>>> [1] grid      tools     stats     graphics  grDevices utils
>>>> datasets  methods   base
>>>>
>>>> other attached packages:
>>>> [1] ggplot2_0.8.7   digest_0.4.1    reshape_0.8.3   plyr_0.1.9
>>>> proto_0.3-8     gridExtra_0.5   lattice_0.17-26 gtools_2.6.1
>>>>
>>>> ______________________________________________
>>>> R-help at r-project.org mailing list
>>>> 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.
>>>>
>>>
>>>
>>> --
>>> Felix Andrews / 安福立
>>> Postdoctoral Fellow
>>> Integrated Catchment Assessment and Management (iCAM) Centre
>>> Fenner School of Environment and Society [Bldg 48a]
>>> The Australian National University
>>> Canberra ACT 0200 Australia
>>> M: +61 410 400 963
>>> T: + 61 2 6125 4670
>>> E: felix.andrews at anu.edu.au
>>> CRICOS Provider No. 00120C
>>> --
>>> http://www.neurofractal.org/felix/
>>>
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> 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.
>
> --
> Dr Paul Murrell
> Department of Statistics
> The University of Auckland
> Private Bag 92019
> Auckland
> New Zealand
> 64 9 3737599 x85392
> paul at stat.auckland.ac.nz
> http://www.stat.auckland.ac.nz/~paul/
>