formatting of labels in legends from trans_format

[Thomas]

Hi everyone

I'm having trouble formatting the labels in legends when inversing a transformation. As an example, consider the following:

test <- data.frame(expand.grid(1:10, 1:10), value=runif(100, min=1, max=1000))

p <- qplot(Var1, Var2, data=test, geom='tile', fill=value)

If I where to log transform the fill scale and furthermore do something to the labels and breaks, I would add

p + scale_fill_gradient(trans='log10', guide=guide_colorbar(), breaks=trans_breaks('log10', function(x) 10^x), labels=trans_format('log10', math_format()))

The problem is that the labels are plotted as unevaluated calls and thus not very pretty... I have tried different permutations of using eval(), parse() and expression() around trans_format() without luck... Is it not possible to put expression as legend labels when dealing with scale transformations?

best wishes

Thomas

[Brandon Hurr]

I was playing and it looks like the problem is as.name(). It puts single backwards quotes around the number which do not get parsed. When I took it out it seemed to work ok... check attachment. 

>math_format

function (expr = 10^.x, format = force) 

{

    quoted <- substitute(expr)

    subs <- function(x) {

        do.call("substitute", list(quoted, list(.x = as.name(x))))

    }

    function(x) {

        x <- format(x)

        llply(x, subs)

    }

}

<environment: namespace:scales>

[Thomas]

Thats definitely an improvement, but I had hoped it was possible to plot actual expressions, as it is with position scales... Is this not possible for legend labels?

best

Thomas

[Brandon Hurr]

Yes, sorry, I realize what you're saying now. It's still not parsing the ^ into an exponent. If you look at the guts, I don't see where it's being told to parse it as an expression at all. 

> tf <- trans_format("log10", math_format())
> tf(10 ^ 1:6)
[[1]]
10^`1`

[[2]]
10^`0.9542425`

[[3]]
10^`0.90309`

[[4]]
10^`0.845098`

[[5]]
10^`0.7781513`

> trans_format
function (trans, format = scientific_format())
{
   if (is.character(trans))
       trans <- match.fun(trans)
   function(x) {
       x <- trans(x)
       format(x)
   }
}
<environment: namespace:scales>

> math_format
function (expr = 10^.x, format = force)
{
   quoted <- substitute(expr)
   subs <- function(x) {
       do.call("substitute", list(quoted, list(.x = as.name(x))))
   }
   function(x) {
       x <- format(x)
       llply(x, subs)
   }
}
<environment: namespace:scales>

> --
> You received this message because you are subscribed to the ggplot2 mailing list.
> Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
>  
> To post: email ggp...@googlegroups.com
> To unsubscribe: email ggplot2+u...@googlegroups.com
> More options: http://groups.google.com/group/ggplot2

[Kohske Takahashi]

This is a bug: https://github.com/hadley/ggplot2/pull/594
Thanks.

kohske

2012/6/26 Brandon Hurr <brando...@gmail.com>:

--
Kohske Takahashi <takahash...@gmail.com>

Assistant Professor,
Research Center for Advanced Science and Technology,
The University of  Tokyo, Japan.
http://www.fennel.rcast.u-tokyo.ac.jp/profilee_ktakahashi.html