These functions save a transcript of your commands and their output to a script file, possibly for later processing with the "enscript" or "pandoc" program.
They work as a combinations of sink
and history
with a
couple extra bells and whistles.
txtStart(file, commands=TRUE, results=TRUE, append=FALSE, cmdfile,
visible.only=TRUE)txtStop()
txtComment(txt,cmdtxt)
txtSkip(expr)
etxtStart(dir = tempfile("etxt"), file = "transcript.txt",
commands = TRUE, results = TRUE, append = FALSE,
cmdbg = "white", cmdcol = "red", resbg = "white",
rescol = "navy", combg = "cyan", comcol = "black",
cmdfile, visible.only = TRUE)
etxtStop()
etxtComment(txt, cmdtxt)
etxtSkip(expr)
etxtPlot(file=paste(tempfile('plot',R2txt.vars$dir),'.eps',sep=''),
width=4, height=4)
wdtxtStart(commands=TRUE, results=TRUE, fontsize=9, cmdfile,
visible.only=TRUE)
wdtxtStop()
wdtxtComment(txt,cmdtxt)
wdtxtSkip(expr)
wdtxtPlot(height=5, width=5, pointsize=10)
mdtxtStart(dir=tempfile('mdtxt'), file='transcript.md',
commands=TRUE, results=TRUE, append=FALSE,
cmdfile, visible.only=TRUE)
mdtxtStop()
mdtxtComment(txt,cmdtxt)
mdtxtSkip(expr)
mdtxtPlot(file=tempfile('plot',R2txt.vars$dir,'.png'),
width=4, height=4)
Most of these commands do not return anything of use. The exceptions are:
etxtStop
returns the name of the transcript file (including the
directory path).
txtSkip
, etxtSkip
, wdtxtSkip
, and
mdtxtSkip
return the value of expr
.
Directory to store transcript file and any graphics file in
Text file to save transcript in
Logical, should the commands be echoed to the transcript file
Logical, should the results be saved in the transcript file
Logical, should we append to file
or replace it
Background color for command lines in file
Color of text for command lines in file
Background color for results sections in file
Text color of results sections in file
Background color for comments in file
Text color of comments in file
A filename to store commands such that it can be
source
d or copied and pasted from
Should non-printed output be included, not currently implemented.
Text of a comment to be inserted into file
Text of a comment to be inserted into cmdfile
An expression to be executed without being included in
file
or cmdfile
Width of plot, passed to dev.copy2eps
,
wdPlot
, or dev.copy
Height of plot, passed to dev.copy2eps
,
wdPlot
, or dev.copy
Size of font to use in MSWord
passed to wdPlot
Greg Snow, 538280@gmail.com
These functions are used to create transcript/command files of your R session. There are 4 sets of functions, those starting with "txt",those starting with "etxt", and those starting with "wdtxt" and those starting with "mdtxt". The "txt" functions create a plain text transcript while the "etxt" functions create a text file with extra escapes and commands so that it can be post processed with enscript (an external program) to create a postscript file and can include graphics as well. The postscript file can be converted to pdf or other format file.
The "wdtxt" functions will insert the commands and results into a Microsoft Word document.
The "mdtxt" functions create a text file but with MarkDown escapes so that it can be post processed with "pandoc" (an external program) to create other formats such as html, pdf, MS Word documents, etc. If the command starts with the string "pander" or "pandoc" (after optional whitespace) then the results will be inserted directly, without escapes, into the transcript file. This assumes that you are using code from the "pander" package which generates markdown formatted output. This will create nicer looking tables and other output.
If results
is TRUE and commands
is FALSE then the result
is similar to the results of sink
. If commands
is true as well then the transcript file will show both the commands
and results similar to the output on the screen. If both
commands
and results
are FALSE then pretty much
the only thing these functions will accomplish is to waste some
computing time.
If cmdfile
is
specified then an additional file is created with the commands
used (similar to the history
command), this file can be
used with source
or copied and
pasted to the terminal.
The Start functions specify the file/directory to create and
start the transcript, wdtxtStart
will open Word if it is
not already open or create a connection to an open word window.
The prompts are changed to remind you
that the commands/results are being copied to the transcript.
The Stop functions stop the recording and reset the prompts.
The R parser strips comments and does some reformatting so the
transcript file may not match exactly with the terminal
output. Use the txtComment
, etxtComment
,
wdtxtComment
, or mdtxtComment
functions to add a comment. This will show up as a line offset
by whitespace in the transcript file, highlighted in the
etxt version, and the default font in Word.
If cmdtxt
is specified then that line
will be inserted into cmdfile
preceded by a # so it
will be skipped if sourced or copied.
The txtSkip
, etxtSkip
, wdtxtSkip
, and
mdtxtSkip
functions will run the code in expr
but
will not include the commands or results in the transcript file
(this can be used for side computations, or requests for help,
etc.).
The etxtPlot
function calls dev.copy2eps
to
create a copy of the current plot and inserts the proper command
into the transcript file so that the eps file will be included
in the final postscript file after processing.
The wdtxtPlot
function calls wdPlot
to send a
copy of the current graph to MS Word.
The mdtxtPlot
function calls dev.copy
to create a
copy of the current plot as a .png file and inserts the proper
command into the transcript file so that the .png file will be
included when processing with pandoc.
if (FALSE) {
etxtStart()
etxtComment('This is todays transcript')
date()
x <- rnorm(25)
summary(x)
stem(x)
etxtSkip(?hist)
hist(x)
etxtPlot()
Sys.Date()
Sys.time()
my.file <- etxtStop()
# assumes enscript and ps2pdf are on your path
system(paste('enscript -e -B -p transcript.ps ', my.file) )
system('ps2pdf transcript.ps')
# if the above commands used mdtxt instead of etxt and the pandoc
# program is installed and on your path (and dependent programs) then use:
system(paste('pandoc -o transcript.docx ', my.file))
}
Run the code above in your browser using DataLab