[R] Gridbase basic question
Paul Murrell
p.murrell at auckland.ac.nz
Tue Sep 28 22:25:50 CEST 2004
Hi
Sean Davis wrote:
> Paul,
>
> Thanks for the extensive and clear explanation. The reason I started
> with grid is that I am hoping to use a combination of segments,
> rectangles, and text to describe relatively complex (genes) objects
> that relate to the x-axis in a plot. I do appreciate the insight from
> you and others that the added complexity of grid may not be necessary.
One alternative possibly not yet explored is that you could just use
grid. If all you need the traditional graphics for is a bounding box
and some axes, then something like ...
x <- sample(1:10, 10)
y <- 1:10
w <- runif(10)
h <- 0.5
pushViewport(plotViewport(c(5, 4, 4, 2)), dataViewport(x, y))
grid.rect()
grid.xaxis()
grid.yaxis()
grid.rect(x=x, y=y, width=w, height=h, default.units="native")
popViewport(2)
... does the trick (again, the advantage is that you don't need to mess
with gridBase).
Paul
> On Sep 27, 2004, at 5:57 PM, Paul Murrell wrote:
>
>> Hi
>>
>>
>> Sean Davis wrote:
>>
>>> All,
>>> I have a simple plot(x,y) and I would like to then insert rectangles
>>> of some length (in native coordinates) and height fixed to 0.5 in
>>> native coordinates. I can't quite get the code right to do this.
>>> Can anyone give me a quick example of how to do this? I looked the
>>> gridBase index and the tutorial (from R-news?) but just haven't
>>> gotten it down yet.
>>> > plot(1:10,1:10)
>>> > par(new=T);vps <- baseViewports()
>>> > pushViewport(vps$inner,vps$figure,vps$plot)
>>> viewport[GRID.VP.28]
>>
>>
>>
>> At this point you are within a viewport which has x- and y-scales
>> corresponding to the plot(1:10, 1:10).
>>
>>
>>> > pushViewport(viewport(x=unit(1,"native"),y=unit(2,"native")))
>>> viewport[GRID.VP.29]
>>
>>
>>
>> You have just created a new viewport at location (1, 2) in the plot,
>> but the scales on this new viewport are the default (0, 1). i.e.,
>> you are now in a completely different coordinate system. Also, this
>> new viewport is as wide and as high as the plot region -- for
>> example, it extends well beyong the left edge of the window/page.
>>
>>
>>> grid.rect(height=unit(0.5,"native"),width=unit(1.5,"native"),just='bot
>>> to m')
>>
>>
>>
>> This draws a rectangle half as high as the current viewport and 1.5
>> times as wide (the native scale in the current viewport is (0, 1) in
>> both dimensions). Importantly, the "native" coordinate systems you
>> are referring to no longer correspond to the scales on the plot.
>>
>>
>>> This draws a very large rectangle going from 2 to 7 (y) and to 8 (x).
>>
>>
>>
>> Three things:
>>
>> (i) If drawing rectangles relative to the current "native" (or user)
>> coordinates is all you want to do then you could just use rect() and
>> ignore gridBase altogether. For example, ...
>>
>> x <- sample(1:10, 10)
>> y <- 1:10
>> w <- runif(10)
>> h <- 0.5
>>
>> plot(1:10,1:10)
>> rect(x - w/2, y - h/2, x + w/2, y + h/2)
>>
>>
>> (ii) Using grid and gridBase, the above example becomes ...
>>
>> plot(1:10,1:10)
>> par(new=T);vps <- baseViewports()
>> pushViewport(vps$inner,vps$figure,vps$plot)
>> grid.rect(x=x, y=y, width=w, height=h, default.units="native")
>> popViewport(3)
>>
>> ... but as mentioned, this is like using a sledge hammer to kill a
>> cat or whatever the expression is.
>>
>> (iii) There would be justification in using grid and gridBase if you
>> want to draw more than just a rectangle, especially if you want to
>> use coordinates other than native. Here's a trivial example (adds
>> fixed size "whiskers" to the corners of the rectangles) ...
>>
>> plot(1:10,1:10)
>> par(new=T);vps <- baseViewports()
>> pushViewport(vps$inner,vps$figure,vps$plot)
>> for (i in 1:10) {
>> pushViewport(viewport(x=x[i], y=y[i], width=w[i], height=h,
>> default.units="native"))
>> grid.rect()
>> grid.segments(0, 0, unit(-1, "mm"), unit(-1, "mm"))
>> grid.segments(0, 1, unit(-1, "mm"),
>> unit(1, "npc") + unit(1, "mm"))
>> grid.segments(1, 1,
>> unit(1, "npc") + unit(1, "mm"),
>> unit(1, "npc") + unit(1, "mm"))
>> grid.segments(1, 0,
>> unit(1, "npc") + unit(1, "mm"),
>> unit(-1, "mm"))
>> popViewport()
>> }
>>
>> ... (but pushing a viewport per data point like this is a LOT slower).
>>
>> Hope that helps
>>
>> Paul
>> --
>> 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/
>
--
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/
More information about the R-help
mailing list