##------------------------------------------------------------------------------#
##   Statistical Analysis of the PHALANX Project.                               #
##                                                                              #
##   MAS in Medizinphysik                                                       #
##   Blockkurs "Computer in der Medizin"                                        #
##------------------------------------------------------------------------------#

##------- Short version of ./PHALANX_statistics.R -----------------------


##------------------------------------------------------------------------------#
## Directories and filenames                                                    #
##------------------------------------------------------------------------------#

dir.base  <- "D:/Martin/Documents/NDS/Computerkurs_2007/Statistik/R"

## MM:
dir.base  <- "/u/maechler/R/Meetings-Kurse-etc/2007-Biomechanik/R"
dir.csv <- file.path(dir.base, "csv")

## Students:
dir.base  <- "http://stat.ethz.ch/Teaching/maechler/Phalanx"
dir.csv   <- dir.base

if(FALSE) {
    ## Students can  ``bootstrap'' themselves {get the data} via
    FF <- tempfile()
    download.file("http://stat.ethz.ch/Teaching/maechler/Phalanx/read-etc.R", FF)
    source(FF)
}

##------------------------------------------------------------------------------#
## Read and merge data.                                                         #
##------------------------------------------------------------------------------#

## Simple utility function:
readTabPhalanx <- function(filename) {
    head1 <- read.table(filename, nrows = 1, sep = ";", as.is = TRUE)
    dat <- read.table(filename, skip = 2,  sep = ";", col.names = head1)
    stopifnot(names(dat)[1] == "SPECIMEN")
    ## use "SPECIMEN" as row names and drop as variable :
    rownames(dat) <- as.character(dat[,"SPECIMEN"]) # and not factor
    ## return
    dat[, -1] # '-1' : leave away first variable  i.e. 'SPECIMEN'
}

### -------------- Now read the 4 files ----------------------------------

d.anthro <- readTabPhalanx(file.path(dir.csv, "PHALANX_Anthropometric.csv"))
d.dist   <- readTabPhalanx(file.path(dir.csv, "PHALANX_Distal_25percent.csv"))
d.mech   <- readTabPhalanx(file.path(dir.csv, "PHALANX_Mechanical_Testing.csv"))
d.spher  <- readTabPhalanx(file.path(dir.csv, "PHALANX_Spherical_VOI.csv"))

## merge

rnMerge <- function(d1,d2, ...) {
    ## Purpose: merge by row.names -- and make the "row.names" variable to row.names
    ## ----------------------------------------------------------------------
    ## Arguments: d1, d2: two data frames with common row names
    ## ----------------------------------------------------------------------
    ## Author: Martin Maechler, Date: 13 Mar 2007

    d <- merge(d1, d2, by = "row.names", ...)
    ## set row names and omit them as variable:
    rownames(d) <- d[, 1]
    d[, -1]
}

d.A.M <- rnMerge(d.anthro, d.mech)

## Martin Stauber's :
d.distal    <- rnMerge(d.A.M, d.dist)
d.spherical <- rnMerge(d.A.M, d.spher)
## All of them:
d.All <- rnMerge(d.distal, d.spher, suffixes= c("_d", "_s"))

## Variabel-Namen vereinfachen: ----------

## Use a simple function :
mySub <- function(nms) {
    sub("Nr.of.", "Nr_", # shorten
        sub("^Ap", "AP", # <<-- Fix typo in file
            gsub("\\.Diameter$", ".dm", # shorten..
                 gsub("\\.$", "",         # omit trailing "."
                      gsub("\\.([._])", "\\1", # substitute ".." by "." and "._" by "_"
                           nms)))))
}

names(d.All) <- mySub(names(d.All))

## All variables but the "mechanical":
d.X <- rnMerge(d.anthro, rnMerge(d.dist, d.spher, suffixes= c("_d", "_s")))
names(d.X) <- mySub(names(d.X))