Downloading and installing the Bwidget library

1  Using widgets from the Bwidget library

One of the handiest widgets available is the "Tree" widget, which shows the upper level of a hierarchy of objects, and which displays the lower levels of the hierarchy when the appropriate graphic element is clicked. Here's an illustration of the tree widget displaying some data frames; the wine data frame's variables are displayed, since the plus sign to the left of its name was clicked.
tree.png
The Bwidget library is not included with the version of Tcl that comes with R under Windows, but it is very easy to install because it is "pure" Tk; in other words, no programs have to be compiled in order to install the library. To install the Bwidget library under Windows, follow these steps:
  1. Go to http://tcl.activestate.com/software/tcllib/
  2. Click on the link labeled "Download distributions"
  3. Click on the "Download" link for "BWidget".
  4. Click on the link labeled "BWidget-1.7.0.zip", and follow the instructions to download this file to your computer.
  5. Find the location of the Tcl libraries that were installed as part of R. This will probably be c:\Program Files\R\R-2.2.1\Tcl\lib, but may vary if you're using a different version of R. In any event, you want to find the lib directory in the Tcl folder.
  6. Extract the files contained in the zip file into the folder you found in the previous step. Once this is done, the tree widget should be available.
Here's the program that uses the tree widget: 
require(tcltk)
tclRequire("BWidget")

doit = function(...){
   use = tclvalue(tkcmd(treeWidget,"selection","get"))
   parts = strsplit(use,':')[[1]]
   df = get(parts[1])
   var = df[[parts[2]]]
   title = paste(parts,compress='$')
   type = tclvalue(type_)
   funs = list(hist = function(x,title)hist(x,main=title),
               box  = function(x,title)boxplot(x,main=title),
               dens = function(x,title)plot(density(x,na.rm=TRUE),type='l',main=title))
   funs[[type]](var,title)
}

main = tktoplevel()
tkwm.title(main,"Tree (Drill-Down) Widget")

dffr = tkframe(main)

xScr       = tkscrollbar(dffr,command=function(...)tkxview(treeWidget,...),orient="horizontal")
yScr       = tkscrollbar(dffr,command=function(...)tkyview(treeWidget,...))
treeWidget = tkwidget(dffr,"Tree",xscrollcommand=function(...)tkset(xScr,...),
                                 yscrollcommand=function(...)tkset(yScr,...),width=30,height=15)

tkpack(yScr,side='right',fill='y')
tkpack(xScr,side='bottom',fill='x')
tkpack(treeWidget,fill='both',expand=1)
tkpack(dffr)

oo = objects()
use = oo[sapply(oo,function(x)class(get(x))) == 'data.frame']

for(i in use){
	tkinsert(treeWidget,"end","root",i,text=i)
        nms = names(get(i))
	classes = sapply(get(i),class)
        modes = sapply(get(i),mode)
        wh = (classes != 'factor' & modes == 'numeric')
        y = nms[wh]
	for(j in y)tkinsert(treeWidget,"end",i,paste(i,j,sep=":"),text=j)
}

chfr = tkframe(main)

type_ = tclVar()

f1 = tkframe(chfr)
tkpack(tklabel(f1,text='Hist',width=8),side='left')
tkpack(tkradiobutton(f1,variable=type_,value='hist'),side='left')

f2 = tkframe(chfr)
tkpack(tklabel(f2,text='Box',width=8),side='left')
tkpack(tkradiobutton(f2,variable=type_,value='box'),side='left')

f3 = tkframe(chfr)
tkpack(tklabel(f3,text='Dens',width=8),side='left')
tkpack(tkradiobutton(f3,variable=type_,value='dens'),side='left')

tkpack(f1,side='left')
tkpack(f2,side='left')
tkpack(f3,side='left')

tkpack(chfr,side='top')

tkpack(tkbutton(main,command=doit,text='Plot'),side='bottom')
When elements are inserted in the tree with tkinsert, three additional arguments have to be provided. The first tells where the item will be inserted. Using a value of ënd" is usually appropriate. The second represents an identifier for the place within the hierarchy that this item should be inserted. If it's a toplevel entry (like the data frame names in this example), the "root" identifier is used. The third argument represents the identifier for the element being inserted. All of these identifiers must be unique, and it is this third argument that will be returned to identify which element of the tree has been selected. In this example, I use the data frame names as the toplevel components, and the data frame name and variable name separated by a colon as the lower level identifiers.


File translated from TEX by TTH, version 3.67.
On 3 May 2006, 12:13.