R help - Jan 2009 - proto question

Dear R Users,

 

I have a couple of proto objects like:

 

wedge <- proto(expr={

    start.year <- 2008

    end.year <- 2050

})

 

star.rating <- wedge$proto(

    star = c(4, 5, 8, 10),

    gain = c(0, .3, .5, .7),

    cost = c(0, 2100, 4000, 7500),

    

    star.rating <- function(., year) 6.0,

 

    setup = function(.){

        .$cost.for.star <- approxfun(.$star, .$cost)

        .$gain.for.star <- approxfun(.$star, .$gain)

    },

 

    test = function(., year) {

        gs <- .$with(gain.for.star)(.$star.rating(year))

    }

)

 

 

 

And a function to create and modify 'instances':

 

create.star.rating <- function(switch.years, star.ratings) {

    res <- star.rating$proto(switch.years = switch.years, star.ratings
star.ratings)

    res$setup()

    res

}

 

 

When I use them as follows:

ee <- create.star.rating(ratings, switch.years)

ee$test(2009)

 

 

The second line gives me the error message:

"Error in .$with(gain.for.star)(.$star.rating(year)) : 

   object "y" not found"

 

I understand why this happens (when objects are inserted into a proto
their environment is set to that of the proto)... but I can't figure out
how to get around it!

 

I have tried setting the environments of the functions created in setup
as follows:

 

setup = function(.){

        t1 <- approxfun(.$star, .$cost)

        t2 <- approxfun(.$star, .$gain)

        .$cost.for.star <- t1

        .$gain.for.star <- t2

        environment(.$cost.for.star) <- environment(t1)

        environment(.$gain.for.star) <- environment(t2)

    }

 

 

But then I get the error:

"Error in get("gain.for.star", env = `*tmp*`, inherits = TRUE) : 

  object "*tmp*" not found"

 

 

Which I think occurs because the last two lines in 'setup' only create
references to their respective environments which disappear on exit of
'setup'.

 

 

 

Any ideas?

 

Regards,

Simon Knapp.

 


	[[alternative HTML version deleted]]

On Wed, Jan 7, 2009 at 10:22 AM, Simon Knapp
<S.Knapp at mmassociates.com.au> wrote:
> Dear R Users,
>
>
>
> I have a couple of proto objects like:
>
>
>
> wedge <- proto(expr={
>
>    start.year <- 2008
>
>    end.year <- 2050
>
> })
>
>
>
> star.rating <- wedge$proto(
>
>    star = c(4, 5, 8, 10),
>
>    gain = c(0, .3, .5, .7),
>
>    cost = c(0, 2100, 4000, 7500),
>
>
>
>    star.rating <- function(., year) 6.0,
>
>
>
>    setup = function(.){
>
>        .$cost.for.star <- approxfun(.$star, .$cost)
>
>        .$gain.for.star <- approxfun(.$star, .$gain)
>
>    },
>
>
>
>    test = function(., year) {
>
>        gs <- .$with(gain.for.star)(.$star.rating(year))
>
>    }
>
> )
>
>
>
>
>
>
>
> And a function to create and modify 'instances':
>
>
>
> create.star.rating <- function(switch.years, star.ratings) {
>
>    res <- star.rating$proto(switch.years = switch.years, star.ratings
> star.ratings)
>
>    res$setup()
>
>    res
>
> }
>
>
>
>
>
> When I use them as follows:
>
> ee <- create.star.rating(ratings, switch.years)
Did you define switch.years somewhere?  The order of
arguments in the call to create.star.rating seems
reversed from the definition.  There may be other problems
too but suggest you start by fixing those.
>
> ee$test(2009)
>
>
>
>
>
> The second line gives me the error message:
>
> "Error in .$with(gain.for.star)(.$star.rating(year)) :
>
>   object "y" not found"
>
>
>
> I understand why this happens (when objects are inserted into a proto
> their environment is set to that of the proto)... but I can't figure
out
> how to get around it!
>
>
>
> I have tried setting the environments of the functions created in setup
> as follows:
>
>
>
> setup = function(.){
>
>        t1 <- approxfun(.$star, .$cost)
>
>        t2 <- approxfun(.$star, .$gain)
>
>        .$cost.for.star <- t1
>
>        .$gain.for.star <- t2
>
>        environment(.$cost.for.star) <- environment(t1)
>
>        environment(.$gain.for.star) <- environment(t2)
>
>    }
>
>
>
>
>
> But then I get the error:
>
> "Error in get("gain.for.star", env = `*tmp*`, inherits =
TRUE) :
>
>  object "*tmp*" not found"
>
>
>
>
>
> Which I think occurs because the last two lines in 'setup' only
create
> references to their respective environments which disappear on exit of
> 'setup'.
>
>
>
>
>
>
>
> Any ideas?
>
>
>
> Regards,
>
> Simon Knapp.