I'm looking to streamline my Sweave document creation, and I'd like to hear about people's current setups. I feel like the holy grail goes something like this:
Editing Rnw code on one half of the
screen
Single keybinding compiles
Sweave document and runs pdflatex
View PDF
on the other half of the screen; once
compiled, PDF is refreshed and centered around the portion of the document you're editing
If compilation has errors, replace the PDF with the results of the compilation (e.g. latex errors or Sweave errors)
I am guessing/hoping that the solution is part Emacs/ESS combined with some code for the Emacs profile and/or a nice Makefile. But I would really like to hear about everybody's preferred way of creating Sweave and/or Latex documents.
A few other R users I talked to use a 'one-directory-per-project' setup, and a simple Makefile. As you suspected, that works well with Emacs/ESS.
I tend to just call a simple shell script sweave which I wrote before before 'R CMD Sweave' was added (as I find re-creating or copying the Makefile unappealing, YMMV). I also use Emacs and an auto-refreshing pdf viewer (like okular or kpdf). Emacs23 can preview pdf files directly too but I have yet to switch my work flow to that.
edd#ron:~$ cat bin/sweave
#!/bin/bash -e
function errorexit () {
echo "Error: $1"
exit 1
}
function filetest () {
if [ ! -f $1 ]; then
errorexit "File $1 not found"
fi
return 0
}
if [ "$#" -lt 1 ]; then
errorexit "Need to specify argument file"
fi
BASENAME=$(basename $1 .Rnw)
RNWFILE=$BASENAME.Rnw
filetest $RNWFILE
echo "library(tools); Sweave(\"$RNWFILE\")" \
| R --no-save --no-restore --slave
LATEXFILE=$BASENAME.tex
filetest $LATEXFILE && pdflatex $LATEXFILE
You can do everything that you suggest there with the StatET plugin for Eclipse. That's what I use for Sweave development; it understands both latex and R very well, including syntax highlighting, etc.
You can get it here: http://www.walware.de/goto/statet.
Longhow Lam has written a nice guide: http://www.splusbook.com/Rintro/R_Eclipse_StatET.pdf.
http://www.statalgo.com/?p=93
I use TeXShop on OS X to produce all of my LaTeX and Sweave reports. For me, a new compilation pipeline is as simple as adding a file, called Sweave.engine to ~/Library/TeXShop/Engines/ which contains the following:
#!/usr/bin/env Rscript
args <- commandArgs(T)
fname <- strsplit(args[1],'\\.')[[1]][2]
Sweave(paste(fname,'Rnw',sep='.'))
system(paste('pdflatex',paste(fname,'tex',sep='.')))
Sweave is now a selectable method of compiling a document inside TeXShop. I can set it to be the default for a document by adding the following TeX hash-bang to the top of the file:
% !TEX TS-program = Sweave
Hitting Cmd-T will typeset the document- the pdf automatically pops up in a separate window. TeXShop also incorporates SyncTeX technology so a Cmd-Click in the Rnw source will highlight the corresponding output in the PDF window and a Cmd-Click in the PDF window will highlight the corresponding input in the Rnw source.
TeXShop is mac-only but a great Qt/poppler-based clone, TeXworks, is available for Linux, Windows and Mac and supports many of the same features-- including TeX hash-bangs and SyncTeX. TeXworks has reached a level of maturity where it is included in version 2.8 of the MikTeX package for Windows.
Try RStudio.
I have been a fan of Emacs and TeXShop as mentioned in previous answers.
However, Rstudio is starting to win me over. It is a rapidly-improving dedicated IDE for R. Definitely worth checking out.
I still love doing certain R-only development tasks in the standard R IDE for mac. But for Sweave documents and some associated R devel at the same time, RStudio wins. It works with virtually zero tweaking. I'm not sure about the PDF-related features in the latter half of the original question.
One thing that has saved me some time is the 'auto-insert' mode in emacs. I have it set up so that each time I open a new .rnw file, emacs automatically sets up a basic document template and all I need to do is start writing my report.
Update: I've switched away from auto-insert. Now I use the "template.el" approach.
I use the "one-directory-per-project" and Makefile approach as well. I also include commands to create output in HTML, which can then be converted to OOo and MS Word, using tth. This is important for me since a lot of my collaborators are MS Office users and are resistant to using the PDF output. I learned a lot about how to do this from Frank Harrell's twiki at Vanderbilt.
Personally I use gvim as my editor of choice and running make from there is quite simple, as it is from Emacs.
Related
This might seem to be a very very stupid question.
But for the past few years I have been using vim and cscope on the terminal, with some screen to make life a bit bearable.
I have just started to learn emacs and it is much more satisfying to use it.
Problem using emacs:
Every time I do M-x find-c-symbol, I get a new buffer with a bunch of files, but I don't know how to open the file at the exact line number.
I googled a lot and found this to open file under cursor: M-x ffap
but this opens at the first line. Can some emacs expert help me??
Thanks
I've got both Emacs 23.x and 24.x installed, both setup with working cscope and xcscope installs. Neither has a "find-c-symbol" function, but there is a "cscope-find-this-symbol", which is what I assume you're actually using.
I'm going to assume you're using a GUI version, and not the text-only version, and that you're actually getting the *cscope* buffer automatically being opened and created (since that's what it sounds like from your description).
For a simple search, I'll get results that look like this:
Finding symbol: debug
Database directory: /home/user/emacs_tags/modular/
-------------------------------------------------------------------------------
*** /home/user/code/modular/frontend/common/controller.test/src/MainTest.cpp:
<global>[73] #ifdef debug
-------------------------------------------------------------------------------
Database directory: /home/user/emacs_tags/rrsdk/
-------------------------------------------------------------------------------
*** /home/user/code/rrsdk/fs/apps/busybox/src/shell/ash.c:
<global>[303] #define debug optlist[15 + ENABLE_ASH_BASH_COMPAT]
*** /home/user/code/rrsdk/bootloader/u-boot/src/board/mcc200/auto_update.c:
<global>[53] #undef debug
<global>[55] #define debug(fmt,args...) printf (fmt ,##args)
<global>[57] #define debug(fmt,args...)
-------------------------------------------------------------------------------
Search complete. Search time = 22.44 seconds.
Assuming your results look similar (they should), there are two multiple target areas in the result. Each file line (the lines starting with ***) is a target to the start of that file. Each individual result is also a target area. If you click on one of the lines that lists a specific match (or put your cursor on it and press enter), it will attempt to jump to the specific line matching the result. If it's not jumping to the specific line correctly it usually means your code has changed since the last time the cscope index file was generated.
I'm not sure how you're using the cscope tool, but you can setup xcscope to auto-index on every change to keep the file up to date, but it really only works for smaller code bases where you can keep the cscope.out files in the top level directory and provide it with a full file list for the files to index. Most people I've talked to use the cscope tool by hand in an external script to manually index/re-index every once in a while and then just interface to the existing cscope database(s) using the emacs tools (mine takes about 4 hours to generate the cscope database for a project that includes the Linux kernel as a sub-part).
I want to be able to send entire files and selections to the SublimeREPL and eval them, but it doesn't work. CoffeeScript REPL will only read one line at a time.
Coffee will only take single lines, until you use ctrl+v to switch CoffeeScript to multiline (which works in Terminal), but is blocked by sublime's interface... I'd really like to be able to eval more than one line using SublimeREPL's "ctrl+, s", "ctrl+, f" key bindings. There seems to be no direct way to open coffee with multiline mode automatically.
Maybe I'm going about this wrong, but it's only reasonable that I could use multiline JS Object notation in a CoffeeScript file, for example, and then test it in the REPL. I do it in Python and straight up Node all the time.
You need to be able to switch multi-line on then off again, so ideally if I could edit Sublime REPL to run that key shortcut ( ctrl+v ) before and after running a snippet of code that would be the best.
SublimeREPL v2.0.9 will fix your problem using multiline hack in repl.coffee.
[Ctrl+, s] and [ctrl+, f] should work as expected now :)
Release notes: https://github.com/wuub/SublimeREPL/releases/tag/2.0.9
repl.coffee http://coffeescript.org/documentation/docs/repl.html#section-2
I made a plugin: https://github.com/billymoon/Sublime-Pipe-Dream/ which aims to do this - and much much more. It needs patching up a bit - so any help appreciated. I use it on my mac with sublime text 2 to pass arbitrary selected text, or whole buffers (not saved files) to shell commands and either display the result in the console, or replace the selected text with the console output.
This is very handy for converting coffee to js, js to coffee, executing js/coffee/sql, beautifying and uglifying code and anything else where you want to pass text through a command line script.
Context
Have some transcript files from terminal interaction, obtained using traditional Unix command "script".
Those transcripts contain lots of control character (like backspace when editing shell commands), and lots of color code sequences as result of running various commands. Occasionally, even colorful full-terminal (ncurses-based) applications like "emacs -nw" or "aptitude" were run.
At program runtime, TERM environment variable was set as "xterm".
Need 1: read (more or less solved)
I need to read those files again and sometimes copy-paste some small parts.
The trouble is : while one color change here and there is not so much of a problem, their actual density makes the output barely readable. Worse, edited command lines (with cursor jumps and edited words) are completely unreadable.
"Okay" solution
Browse through files using e.g. "less -r". Paging forward in the same terminal setup reproduces the various color and character style.
But many other features come out more or less broken, e.g. search backwards produces jumbled terminal output, often have to pres "Ctrl-L" to clean thing up.
Need 2: editing
My preferred editor is emacs. Some people have had a similar situation when running the shell inside emacs, e.g. Something wrong with Emacs shell.
Here is not the same situation. Examples of differences: here we don't have to run an actual shell, but we need to move cursor freely like in usual editing.
Editing here means easily open such a transcript file in editor and then:
at all times through editing, see character changes (color, attributes) as conveyed by the terminal codes
(optional) some character that are neighbour on the terminal grid but are separated in file by some control characters would have a visual hint about this
ability to insert some text,
delete sections,
use all editor features like search/replace etc.
copy & paste to & from file (including to external programs, which would receive just plain text)
in my wildest dreams, some kind of "flatten" action, like select a sequence with a heavily edited command line and replace it with a simple series of characters as if it was typed in one run. "Visual hints" mentioned above would disappear.
Type Alt-: to evaluate something in the minibuffer. Evaluate (ansi-color-apply-on-region (point-min) (point-max)) and it will convert ansi color codes to be font colors.
I am maintaining someone else's code. The code is written in C using GCC 4.4.3 on Linux platform. However, the code jumps around a lot and its difficult to find out where all the functions are called from.
In Visual Studio, there is a feature called 'Call Hierarchy' which will display where functions are called from and called to. Does Emacs (23.1.1) have any such feature?
The classic Emacs way to jump to calls is to use TAGS file and use the M-. command. I recommend using Exuberant C Tags with the following command in the root directory of your project :
ctags -e --c-kinds=+pxd -R .
Then using visit-tags-table you can open the TAGS file. With M-. you can jump to each definition or call to your keyword. Use C-u M-. to jump to another occurrence. Use C-x z z z... to repeat the search.
If you have many projects, you can create a TAGS file for each one of them and then call visit-tags-file to add the TAGS file to your list of TAGS files to search from.
Another classic way (the un*x way), is to use the command M-x find-grep to search for occurrences of your keyword.
http://cedet.sourceforge.net/symref.shtml
At my current job, we have coding-style standards that are different from the ones I normally follow. Fortunately, we have a canned RC file for perltidy that I can apply to reformat files before I submit them to our review process.
I have code for emacs that I use to run a command over a buffer and replace the buffer with the output, which I have adapted for this. But I sometimes alternate between emacs and vim, and would like to have the same capabilities there. I'm sure that this or something similar is simple and had been done and re-done many times over. But I've not had much luck finding any examples of vim-script that seem to do what I need. Which is, in essence, to be able to hit a key combo (like Ctrl-F6, what I use in emacs) and have the buffer be reformatted in-place by perltidy. While I'm a comfortable vim-user, I'm completely clueless at writing this sort of thing for vim.
After trying #hobbs answer I noticed that when filtering the entire buffer through perltidy the cursor returned to byte 1, and I had to make a mental note of the original line number so I could go back after :Tidy completed.
So building on #hobbs' and #Ignacio's answers, I added the following to my .vimrc:
"define :Tidy command to run perltidy on visual selection || entire buffer"
command -range=% -nargs=* Tidy <line1>,<line2>!perltidy
"run :Tidy on entire buffer and return cursor to (approximate) original position"
fun DoTidy()
let l = line(".")
let c = col(".")
:Tidy
call cursor(l, c)
endfun
"shortcut for normal mode to run on entire buffer then return to current line"
au Filetype perl nmap <F2> :call DoTidy()<CR>
"shortcut for visual mode to run on the current visual selection"
au Filetype perl vmap <F2> :Tidy<CR>
(closing " added to comments for SO syntax highlighting purposes (not required, but valid vim syntax))
DoTidy() will return the cursor to its original position plus or minus at most X bytes, where X is the number of bytes added/removed by perltidy relative to the original cursor position. But this is fairly trivial as long as you keep things tidy :).
[Vim version: 7.2]
EDIT: Updated DoTidy() to incorporate #mikew's comment for readability and for compatibility with Vim 7.0
My tidy command:
command -range=% -nargs=* Tidy <line1>,<line2>!
\perltidy (your default options go here) <args>
If you use a visual selection or provide a range then it will tidy the selected range, otherwise it will use the whole file. You can put a set of default options (if you have any) at the point where I wrote (your default options go here), but any arguments that you provide to :Tidy will be appended to the perltidy commandline, overriding your defaults. (If you use a .perltidyrc you might not have default args -- that's fine -- but then again you might want to have a default like --profile=vim that sets up defaults only for when you're working in vim. Whatever works.)
The command to filter the entire buffer through an external program is:
:%!command
Put the following in ~/.vimrc to bind it to Ctrl-F6 in normal mode:
:nmap <C-F6> :%!command<CR>
For added fun:
:au Filetype perl nmap <C-F6> :%!command<CR>
This will only map the filter if editing a Perl file.
Taking hobbs' answer a step further, you can map that command to a shortcut key:
command -range=% -nargs=* Tidy <line1>,<line2>!perltidy -q
noremap <C-F6> :Tidy<CR>
And another step further: Only map the command when you're in a Perl buffer (since you probably wouldn't want to run perltidy on any other language):
autocmd BufRead,BufNewFile *.pl,*.plx,*.pm command! -range=% -nargs=* Tidy <line1>,<line2>!perltidy -q
autocmd BufRead,BufNewFile *.pl,*.plx,*.pm noremap <C-F6> :Tidy<CR>
Now you can press Ctrl-F6 without an active selection to format the whole file, or with an active selection to format just that section.
Instead of creating a new keyboard shortcut, how about replacing the meaning of the = command which is already in people's finger memory for indenting stuff? Yes, perlcritic does more than just indent but when you use perlcritic anyways, then you probably don't want to go back to the inferior "just indent" = command. So lets overwrite it!
filetype plugin indent on
autocmd FileType perl setlocal equalprg=perltidy
And now we can use = just like before but with the added functionality of perlcritic that goes beyond just indenting lines:
== run perlcritic on the current line
5== run perlcritic on five lines
=i{ Re-indent the 'inner block', i.e. the contents of the block
=a{ Re-indent 'a block', i.e. block and containing braces
=2a{ Re-indent '2 blocks', i.e. this block and containing block
gg=G run perlcritic on the entire buffer
And the best part is, that you don't have to learn any new shortcuts but can continue using the ones you already used with more power. :)
I'm used to select text using line oriented visual Shift+V and then I press : an I have !perltidy -pbp -et4 somewhere in history so I hit once or more up arrow ⇧.