Henrik Bengtsson HenrikBengtsson

WORK FLOW: Git and Bioconductor SVN (2016)

This is how I currently work with Git and the Bioconductor SVN. I'll use the affxparser package as an example.

Clone existing Git repository

% git clone [email protected]:HenrikBengtsson/affxparser.git
% cd affxparser

Checkout existing Git Flow branches

I'm using the Git Flow workflow everywhere, including this package. In short,

The overhead of method dispatching can be noticable, particularly if done on small objects and it adds up if doing many many times. For instance,

> add <- function(a,b) UseMethod("add")
> add.integer <- function(a,b) a+b

gives with method dispatching:

> system.time({ for (kk in 1:10e6) add(1L, 2L) })
  user  system elapsed
35.43 0.05 35.70

Agilent adds random TABs to end of data rows?!?

Below are the first few lines(*) of two files exported from the Agilent CytoGenomics Software. These files are plain text files with a few "metadata" rows before the actual tab-delimited data begins. The tab-delimited data appears to have a header row with tab-delimited column names followed by corresponding tab-delimited data rows. But it also appears as if there are "random" TABs (ASCII 0x09 = '\t') added to the end of either to header row or the data rows. What's up with that?

PS. (*) I cut out the top few lines by reading the files as a binary file in order to remove the risk of it being me/my editor introducing these TAB outliers. Also, I've verified using binary diff that these example files are identicial to the top of the full files.

Function to count TABs in a string

> countTABs <- function(x) sum(charToRaw(x) == cha

RSP: Including file into Markdown

This can be done using the <%@include file="..."%> RSP preprocessing construct.

From R:

> R.rsp::rfile('main.md.rsp', postprocess=FALSE)
RspFileProduct:
Pathname: C:/Users/hb/braju.com.R/R.rsp/main.md
File size: 66 bytes

	#' Create a temporary file with content
	#'
	#' Create a temporary file with content that will auto delete
	#' as soon as there is no longer any references to it.
	#'
	#' @param content A character string to be written to the file
	#' @param ... Optional arguments to \code{\link[base]{tempfile}()}.
	#'
	#' @return The absolute pathname to the temporary file.
	#'

	## Example showing that vectors/matrices/arrays can indeed have
	## names attributes while still having dimensions/dimnames.
	## I intentionally picked confusing names and dimnames.

	x <- matrix(1:4, nrow=2, dimnames=list(c("a", "b"), c("A", "B")))
	print(x)
	### A B
	### a 1 3
	### b 2 4

	## Identify installed packages that fail to load
	rscript <- function(...) {
	system2(file.path(R.home("bin"), "Rscript"), ...)
	}

	works <- function(pkg) {
	## Need to call in separate R session to not use up all DLLs
	rcmd <- sprintf("cat(isTRUE(requireNamespace('%s')))", pkg)
	as.logical(rscript(args=c("-e", dQuote(rcmd)), stdout=TRUE))
	}

	source("https://gist.githubusercontent.com/HenrikBengtsson/ed1eeb41a1b4d6c43b47/raw/ebe58f76e518dd014423bea466a5c93d2efd3c99/readtable-fix.R")

	kkk <- c("a\tb",
	"3.14\tx")

	colClasses <- c(a="numeric", b="character")
	data <- read.table(textConnection(kkk),
	sep="\t",
	header = TRUE,
	colClasses = colClasses)

	#######################################################################
	# rforge2git.R by Henrik Bengtsson
	#
	# USAGE:
	#
	# Rscript rpkg_svn2git.R <pkg>
	#
	# EXAMPLE:
	# Rscript rpkg_svn2git.R r-forge:r-dots/R.menu
	# Rscript rpkg_svn2git.R bioc-devel:illuminaio

	source("mixedsortRoman.R")

	# Basics
	yi <- c(1, 5, 10, 50, 100)
	print(yi)
	## [1] 1 5 10 50 100

	yr <- as.character(as.roman(yi))
	print(yr)
	## [1] "I" "V" "X" "L" "C"