[R] bootstrap: boot package

[Ben Bolker]

  I've basically done this (although entirely in R rather than using
Perl); see attached.
  (Frank Harrell reported a problem with an earlier version of this, so
I'm attaching the latest.  It seems to work OK on my system with
--vanilla).
  Basically,
  source(...)
  create.index.CRAN()
  help.search.CRAN("foo")

 should work.
  Improvements welcome.

On Mon, 28 Jan 2002, Dirk Eddelbuettel wrote:

> IIRC one problem is help.search() on one's computer is "local", it cannot
> know about packages which are not installed.  So would it be a) feasible and
> b) desirable to create a "meta"-package for nothing but "global"
> documentation searches ? One way would be to go online and query Jon Baron's
> search engine, another to work offline by searching a db-in-a-file of
> keywords stripped from the documentation on a CRAN master [ feeding this
> could be automated with Perl scripts ]
>
> Am I making sense, or do I need more coffee?
>
> Dirk
>

download.index.CRAN <- function(pkg,CRAN=getOption("CRAN"),
                    descrip=paste(srccontrib.url(CRAN),"/Descriptions",sep=""),
                    local.ext=".INDEX.tmp") {
  download.file(paste(descrip,"/",pkg,".INDEX",sep=""),
                paste(pkg,local.ext,sep=""))
}

srccontrib.url <- function(CRAN) {
   paste(CRAN,"/src/contrib",sep="")
}

## these two functions are duplicated from bbfuns package:
##  didn't want to introduce dependence for minor utility functions
## strsplit produces zero-length words if there are leading spaces
nwords <- function(x) {
  sapply(strsplit.words(x),function(z)length(z[nchar(z)>0]))
}

strsplit.words <- function(x) {
  strsplit(gsub("[ \t]*"," ",x)," ")
}

## readline() without the 32-char limit
## [unnecessary with more recent R versions??]
myreadline <- function(prompt="") {
  cat(prompt)
  readLines(n=1)
}

trans.index.CRAN <- function(fn) {
  ##  cat(fn,"\n")
  lines <- scan(fn,what=character(),sep="\n",quiet=TRUE)
  ret <- NULL
  if (length(lines)>0) {
    ## paste continuation lines together
    ## look for initial whitespace; might miss funny formats
    tablines <- grep("^[ \t]",lines)
    lines <- sub("^[\t ]*","",lines) ## now delete initial whitespace
    ## join "continuation sets" (sets of consecutive lines with leading whitespace)
    if (length(tablines)>0) {
      v <- 1:length(lines)
      csets <- v-cumsum(v %in% tablines)
      lines <- sapply(split(lines,csets),paste,collapse=" ")
    }
    pkgname <- gsub("/INDEX","",gsub("[\* ]*","",lines[1]))
    lines <- lines[-1]
    pkg <- rep(pkgname,length(lines))
    fun <- sapply(strsplit.words(lines),"[",1)
    descr <- sapply(strsplit.words(lines),function(z)paste(z[-1],collapse=" "))
    ret <- cbind(pkg,fun,descr)
    dimnames(ret) <- list(NULL,c("Package","Function","Description"))
  }
  ret
}

create.index.CRAN <- function(CRAN=getOption("CRAN"),
                         contriburl=srccontrib.url(CRAN),
                         descrip=paste(srccontrib.url(CRAN),"/Descriptions",sep=""),
                         local.ext=".INDEX.tmp",
                         save=TRUE,savefile="CRAN-index.RData",
                         download=TRUE) {
  if (download) {
    pkglist <- CRAN.packages(contriburl=contriburl)
    sapply(pkglist[,1],download.index.CRAN,CRAN=CRAN,descrip=descrip,
           local.ext=local.ext)
  }
  indfiles <- list.files(pattern=paste("*",local.ext,sep=""))
  CRAN.index <- do.call("rbind",lapply(indfiles,trans.index.CRAN))
  answer <- substr(myreadline("Delete temporary index files (y/N)? "), 1, 1)
  if (answer == "y" | answer == "Y")
    unlink(indfiles, TRUE)
  if (save) save(CRAN.index,file=savefile)
  invisible(CRAN.index)
}

help.search.CRAN <- function(str,indexmat=CRAN.index,ignore.case=TRUE,
                             packages.only=FALSE,
                             load=TRUE,
                             file="CRAN-index.RData") {
  ## check for object/file, load if possible
  if (!exists(deparse(substitute(indexmat)))) {
    if (load) {
      if (!file.exists(file))
        stop("Can't find CRAN index file to load (try create.index.CRAN)")
      load(file)
    }
    else stop ("Can't find CRAN index object (try loading from a file?)")
  }
  ## find matches in either Function or Description fields
  ind <- unique(c(grep(str,indexmat[,"Function"],ignore.case=ignore.case),
                  grep(str,indexmat[,"Description"],ignore.case=ignore.case)))
  if (length(ind)==0)
    cat("No functions found\n")
  else
    if (packages.only) noquote(unique(indexmat[ind,"Package"]))
    else noquote(indexmat[ind,])
}

## junk in index: PTAk (how to filter?)
## give "no functions found" message ?



-- 
318 Carr Hall                                bolker at zoo.ufl.edu
Zoology Department, University of Florida    http://www.zoo.ufl.edu/bolker
Box 118525                                   (ph)  352-392-5697
Gainesville, FL 32611-8525                   (fax) 352-392-3704


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._