R help - Oct 2004 - FW: Plotcorr: colour the ellipses to emphasize the differences

-----Original Message-----
From: Duncan Murdoch [mailto:murdoch at stats.uwo.ca]
Sent: ned 2004-10-17 15:34
To: Gorjanc Gregor
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] Plotcorr: colour the ellipses to emphasize the differences 
 
On Sun, 17 Oct 2004 02:51:58 +0200, "Gorjanc Gregor"
<Gregor.Gorjanc at bfro.uni-lj.si> wrote:

[removed old stuff] 
>>I would also like to know if it is possible to use some other scale of 
>> colors instead of cm.colors, rainbow, heat.colors, terrain.colors, 
>> topo.colors. I would like to have positive correlations in blue and 
>> nagative ones in red spectrum of colors. Is it possible?
> The source of cm.colors is visible (just type "cm.colors" and it
will
> be printed).  You could write your own function to change the scale to
> blue through red instead of cyan through magenta by changing a few
> constants in that function.
Thanks, that helped me to do what I wanted. Thanks. Bellow is the code.

cm.colors <- function (n) 
{
    # from red
    from=0
    # to blue
    to=8/12
    if ((n <- as.integer(n[1])) > 0) {
        even.n <- n%%2 == 0
        k <- n%/%2
        l1 <- k + 1 - even.n
        l2 <- n - k + even.n
        c(if (l1 > 0) 
            hsv(h = from, 
                s = seq(0.5, ifelse(even.n, 0.5/k, 0), length = l1), 
                v = 1), 
          if (l2 > 1) 
            hsv(h = to, 
                s = seq(0, 0.5, length = l2)[-1], 
                v = 1))
    }
    else character(0)
}
> I think it would be useful to have a general function that did what
> cm.colors does but for other paths through colour space, but I've
> never written one (or done a thorough search to see if someone else
> has.)
I agree with this!

With regards,
    Gregor GORJANC

On Sun, 17 Oct 2004 20:33:48 +0200, "Gorjanc Gregor"
<Gregor.Gorjanc at bfro.uni-lj.si> wrote:
>-----Original Message-----
>From: Duncan Murdoch [mailto:murdoch at stats.uwo.ca]
>> I think it would be useful to have a general function that did what
>> cm.colors does but for other paths through colour space, but I've
>> never written one (or done a thorough search to see if someone else
>> has.)
>
>I agree with this!
I've now written a little one:

path.colors <- function(n, path=c('cyan', 'white',
'magenta'),
interp=c('rgb','hsv')) {
    interp <- match.arg(interp)
    path <- col2rgb(path)
    nin <- ncol(path)    
    if (interp == 'hsv') {
	path <- rgb2hsv(path)
	# Modify the interpolation so that the circular nature of hue
is used
	
        for (i in 2:nin) 
	    path[1,i] <- path[1,i] + round(path[1,i-1]-path[1,i]) 

        result <- apply(path, 1, function(x) approx(seq(0, 1,
len=nin), x, seq(0, 1, len=n))$y)
        return(hsv(result[,1] %% 1, result[,2], result[,3]))
    } else {
        result <- apply(path, 1, function(x) approx(seq(0, 1,
len=nin), x, seq(0, 1, len=n))$y)
	return(rgb(result[,1]/255, result[,2]/255, result[,3]/255))
    }
}

I'm not entirely happy with it, but you're welcome to play with it.
The general idea is that you give a sequence of colors in either rgb
or hsv space, and it expands that sequence to a longer string of
interpolated colors.  For example, the default

path.colors(n, path=c("cyan", "white", "magenta"),
interp="rgb")

is fairly close to cm.colors(n) (but more saturated), and

path.colors(n, c("red","green",
"blue","red"), interp="hsv")

is like rainbow(n-1) (but it repeats red at the end).

Duncan Murdoch