shipunov (version 1.13)

hrahn: Angiosperm families: morphological characters

Description

This data originated from the Hansen and Rahn (1969) "punched cards" publication, and subsequent additions and corrections (Hansen and Rahn, 1972; Hansen and Rahn, 1979). Idea was to use paper cards with holes to assist identification of flowering plants (angiosperm) families. These cards were digitized (Duncan and Meacham, 1986) and then used in several multi-entry identification systems (for example, Duncan and Meacham, 1986; Ray, 1995; Families..., 2008).

But what was a sizeable task in 1980--1990s, now is only few hours of R programming. It is therefore quite easy to make such system with R, please see the example. The core function is only a few lines of code, everything else is the interface "bells and whistles". This example system is also applicable to any data with similar structure.

The 'hrahn' data can also be used for the purposes other than identification, for example, to assist in the morphological analysis of angiosperm families.

Comparing with original printed sources, the version used here misses supporting illustrations and some comments to characters. Comparing with digital sources, it was slightly modified, mostly to correct the imperfect digitization, and add some comments from the printed version (they are in lowercase).

One of comments is large but important so it is placed below as "Note I".

===

Note I. [concerning naming of perianth]

A. Perianth segments in 1 cycle or 2 cycles uniform in colour, size and shape.

B. coloured and petal-like ... all _petals_

BB. green (colourless if the plant is without chlorophyll) or dry and hyaline, glumaceous or scarious ... all _sepals_

AA. Perianth segments in 2 cycles different in colour, size or shape.

C. outer cycle ... _sepals_

CC. inner cycle ... _petals_

AAA. Perianth segments spirally arranged with a gradual transition in colour, size and shape from inner to outer segments: in these cases we have guarded against misinterpretations by stating _all_ segments as _sepals_ _and_ as _petals_. If there is a tendency to differentiation into sepals and petals, then the numbers judged by us to be interpretable as sepals are stated as such and in the same way for the petals.

===

The data is based on the family concepts and characters used in Melchior (1964), Hutchinson (1967) and Cronquist (1981). Therefore, family concepts might be different from those which are in use now. In the data, families in are given in accordance with classifications above so outputted list of families is not sorted alphabetically.

Usage

hrahn

Format

This is a list which contains two components:

data

Binary matrix, row names are families, columns with 'chars'

chars

Character vector with descriptions of characters, posititons correspond with columns of 'data'

Examples

# NOT RUN {
data <- hrahn$data
chars <- hrahn$chars

showcharlist <- function(selchar) {
 tmp <- tempfile()
 selected <- ifelse(seq_along(chars) %in% selchar, "[X]", "[ ]")
 useful <- makeuseful(selchar)
 selected[useful] <- "[O]"
 write.table(data.frame(selected, seq_along(chars), chars),
  file=tmp, quote=FALSE, col.names=FALSE, row.names=FALSE)
 file.show(tmp)
}

makeuseful <- function(selchar) { # numbers of potentially useful characters
 selrows <- rowSums(data[, selchar, drop=FALSE]) == length(selchar)
 sums <- colSums(data[selrows, , drop=FALSE])
 seq_len(ncol(data))[sums > 0 & sums < sum(selrows)]
}

makefam <- function(selchar) { # the core function
 selrows <- rowSums(data[, selchar, drop=FALSE]) == length(selchar)
 row.names(data)[selrows]
}

displayfam <- function(selfam, howmany=12) { # display first "howmany" families
 if (is.null(selfam) || length(selfam) == 0) return("None")
 lfam <- length(selfam)
 if (lfam > howmany) {
 dfam <- selfam[seq_len(howmany)]
 res <- paste(c(dfam, paste0("and ", lfam-12, " more")), collapse=", ")
 } else {
 res <- paste(selfam, collapse=", ")
 }
 res
}

updatechar <- function(old, new) { # add or remove characters
 positive <- new[new > 0 & new <= length(chars)]
 old <- union(na.omit(old), positive)
 negative <- abs(new[new < 0])
 setdiff(old, negative)
}

displaydn <- function(num, sym="-") { # display numbers with dashes
 if (!is.numeric(num)) stop("Argument must be numeric")
 if (length(num) == 1) return(as.character(num))
 num <- sort(unique(num))
 if (length(num) == 2) return(paste(num, collapse=", "))
 num[abs(num - c(num[length(num)], num[-length(num)])) == 1 &
  abs(num - c(num[-1], num[1])) == 1] <- "-"
 gsub(", (-, )+", sym, paste(num, collapse=", "))
## slightly longer (but concatenates with +1 number):
## cc <- paste0(num, c(ifelse(diff(num) == 1, "-", ""), ""), collapse=", ")
## gsub("-, ", "-", gsub("-, (-*[0-9]+-, )+", "-", cc))
}

displaychar <- function(selchar) {
 if (is.null(selchar) || length(selchar) == 0) return("None")
 displaydn(selchar)
}

run <- function(howmany=12, selfam=NULL, selchar=NULL) { # interface, recursive function
if (!interactive()) return(cat("Please run in interactive mode\n"))
cat("Results:", displayfam(selfam, howmany=howmany), "\n")
cat("Selected characters:", displaychar(selchar), "\n")
cat("Potentially useful characters:", displaychar(makeuseful(selchar)), "\n")
cat("===\n")
cat("Type (character) numbers, separate with comma, negative numbers remove from selection\n")
cat("Type 'c' to see the list of characters, [X] selected, [O] potentially useful\n")
cat("Type any other single letter to exit\n")
cat("===\n")
x <- readline(prompt="Your choice: ")
while (TRUE) {
 if (x == "c") showcharlist(selchar)
 if (x %in% c(letters[-3], LETTERS)) break
 new <- suppressWarnings(as.integer(strsplit(x, split=",")[[1]]))
 selchar <- updatechar(selchar, new)
 selfam <- makefam(selchar)
 run(howmany=howmany, selfam=selfam, selchar=selchar)
 break
 }
}

run()

# }