R help - Oct 2009 - General means of matching a color specification to an official R color name

Hello List Dwellers:

I?ve looked around quite a bit, but don?t quite see an answer that I
understand.

I?m looking for a way to take any kind of color specification (rgb, hsv,
hcl, hex) and match it to the n-nearest R official color names.  Clearly it
is easy to interconvert different specification schemes and color spaces,
but matching to the name seems a bit trickier.  Seems like if one has a
specification, it could be fuzzy-matched to the list of official R colors
expressed in the same specification.  Unfortunately, I don?t know much about
fuzzy matching.

For example, following some examples I found in the archives and the wiki, I
wrote this little function to create a table of official R colors and sort
it if desired:

colorSpecTable <- function(col = colors(), sort = NULL){
    require(gplots)
    rgbcodes <- t(col2rgb(col))
    names <- col
    hex <- col2hex(col)
    df <- data.frame(name = names, hex.code = hex, rgbcodes)
   # additional elements for other color spaces could be added
    if (!identical(sort, NULL)) df <- sort.data.frame(df, by = sort)
    }

Note that sort.data.frame is from the R-wiki and is appended below.  Is
there a clever way to search a table created by this function, and identify
the n-closest colors based upon some reasonable criteria?  What I hope for
is something like this:

colorMatch <- function(hex = NULL, n, plot = HOPEFULLY) {
    df.rgb <- colorSpecTable(sort = ~red+green+blue) # master table
    # now search for the n closest matches of hex in df.rgb$hex.code
    # perhaps hex should be converted into a different color space 1st
    # eventually would like to display matches side by side w/hex
    }

Thanks as always.  Bryan
*************
Bryan Hanson
Acting Chair
Professor of Chemistry & Biochemistry
DePauw University, Greencastle IN USA

sort.data.frame <- function(x, by){
    # Author: Kevin Wright
    # with some ideas from Andy Liaw
    # Found in the R Wiki

    # x: A data.frame
    # by: A one-sided formula using + for ascending and - for descending
    # Sorting is left to right in the formula
  
    # Useage is:
    # library(nlme);
    # data(Oats)
    # sort(Oats, by= ~nitro-Variety)

    if(by[[1]] != "~")
        stop("Argument 'by' must be a one-sided formula.")

    # Make the formula into character and remove spaces
    formc <- as.character(by[2])
    formc <- gsub(" ", "", formc)
    # If the first character is not + or -, add +
    if(!is.element(substring(formc, 1, 1), c("+", "-")))
        formc <- paste("+", formc, sep = "")

    # Extract the variables from the formula
    vars <- unlist(strsplit(formc, "[\\+\\-]"))
    vars <- vars[vars != ""] # Remove any extra "" terms

    # Build a list of arguments to pass to "order" function
    calllist <- list()
    pos <- 1 # Position of + or -
    for(i in 1:length(vars)){
        varsign <- substring(formc, pos, pos)
        pos <- pos + 1 + nchar(vars[i])
        if(is.factor(x[, vars[i]])){
            if(varsign == "-") {
                calllist[[i]] <- -rank(x[, vars[i]])
            } else {
                calllist[[i]] <- rank(x[, vars[i]])
            }
        } else {
            if(varsign == "-") {
                calllist[[i]] <- -x[, vars[i]]
            } else {
                calllist[[i]] <- x[,vars[i]]
            }
        }
    }
    return(x[do.call("order", calllist), ])
}

On Tue, Oct 13, 2009 at 9:43 PM, Bryan Hanson <hanson at depauw.edu>
wrote:
> Hello List Dwellers:
>
> I?ve looked around quite a bit, but don?t quite see an answer that I
> understand.
>
> I?m looking for a way to take any kind of color specification (rgb, hsv,
> hcl, hex) and match it to the n-nearest R official color names. ?Clearly it
> is easy to interconvert different specification schemes and color spaces,
> but matching to the name seems a bit trickier. ?Seems like if one has a
> specification, it could be fuzzy-matched to the list of official R colors
> expressed in the same specification. ?Unfortunately, I don?t know much
about
> fuzzy matching.
>
> For example, following some examples I found in the archives and the wiki,
I
> wrote this little function to create a table of official R colors and sort
> it if desired:
>
> colorSpecTable <- function(col = colors(), sort = NULL){
> ? ?require(gplots)
> ? ?rgbcodes <- t(col2rgb(col))
> ? ?names <- col
> ? ?hex <- col2hex(col)
> ? ?df <- data.frame(name = names, hex.code = hex, rgbcodes)
> ? # additional elements for other color spaces could be added
> ? ?if (!identical(sort, NULL)) df <- sort.data.frame(df, by = sort)
> ? ?}
>
> Note that sort.data.frame is from the R-wiki and is appended below. ?Is
> there a clever way to search a table created by this function, and identify
> the n-closest colors based upon some reasonable criteria? ?What I hope for
> is something like this:
>
> colorMatch <- function(hex = NULL, n, plot = HOPEFULLY) {
> ? ?df.rgb <- colorSpecTable(sort = ~red+green+blue) # master table
> ? ?# now search for the n closest matches of hex in df.rgb$hex.code
> ? ?# perhaps hex should be converted into a different color space 1st
> ? ?# eventually would like to display matches side by side w/hex
> ? ?}
 You just need to define your distance in colour space. Simplest might
be a euclidean distance in three-dimensional r,g,b coordinates,
something like:

nearColour <- function(r,g,b){
  ctable = col2rgb(colors())
  cdiff = ctable - c(r,g,b)
  cdist = cdiff[1,]*cdiff[1,]+cdiff[2,]*cdiff[2,]+cdiff[3,]*cdiff[3,]
  return(colors()[cdist == min(cdist)])
  }

 This gives colour names nearest to r,g,b triples, with possible
multiple results:

 > nearColour(0,0,0)
 [1] "black" "gray0" "grey0"
 > nearColour(1,1,1)
 [1] "black" "gray0" "grey0"
 > nearColour(255,255,255)
 [1] "white"   "gray100" "grey100"
 > nearColour(128,0,0)
 [1] "darkred" "red4"

Any good? You could also do it in hsv space, but there's probably
enough colours in the colors() vector that it wouldn't make much
difference...

Barry

Hi Bryan,
You can get a near match with the color.id function in the plotrix package.

Jim