20 KiB
Org-babel
- Introduction
- Getting started
- Basic org-babel functionality
- A meta-programming language for org-mode
- Spreadsheet plugins for org-mode in any language
- Library of Babel
- Reproducible Research
- Literate programming
- Reference / Documentation
<div id="subtitle"> <p>executable source code blocks in org-mode</p> </div> <div id="logo"> <p> <img src="images/tower-of-babel.png" alt="images/tower-of-babel.png" title="And the Lord said, Behold, the people is one, and they have all one language; and this they begin to do; and now nothing will be restrained from them, which they have imagined to do. Genesis 11:1-9"/> <div id="attr"> from <a href="http://www.flickr.com/photos/23379658@N05/" title=""><b>Martijn Streefkerk</b></a> </div> </p> </div>
Introduction
Org-babel provides the following modifications to the existing support for blocks of source code examples in the org-mode core.
- source code execution
- arguments to source code blocks
- exportation of source code blocks to files (literate programming)
Getting started
-
Grab the latest code from the git repo at github/org-babel
git clone git://github.com/eschulte/org-babel.git
-
Add the following lines to your .emacs, replacing the path as appropriate. A good place to check that things are up and running would the examples in Basic org-babel functionality.
(add-to-list 'load-path "/path/to/org-babel/lisp") (require 'org-babel-init)
-
Then activate the subset of supported Org-babel languages which you will want to be able to execute on your system. The following can be added to your .emacs and used to activate languages. It includes a brief list of the requirements for each language.
;; Uncomment each of the following require lines if you want org-babel ;; to support that language. Each language has a comment explaining ;; it's dependencies. See the related files in lisp/langs for more ;; detailed explanations of requirements. ;; (require 'org-babel-R) ;; R and ess-mode ;; (require 'org-babel-asymptote) ;; asymptote ;; (require 'org-babel-css) ;; none ;; (require 'org-babel-ditaa) ;; ditaa ;; (require 'org-babel-dot) ;; dot ;; (require 'org-babel-gnuplot) ;; gnuplot, and gnuplot-mode ;; (require 'org-babel-ocaml) ;; ocaml, and tuareg-mode ;; (require 'org-babel-python) ;; python, and python-mode ;; (require 'org-babel-ruby) ;; ruby, irb, ruby-mode, and inf-ruby mode ;; (require 'org-babel-sql) ;; none ;; ;; Once you've activated languages, load the library of babel for ;; pre-built helpers in the languages you will be using. (org-babel-load-library-of-babel)
Basic org-babel functionality
Source code execution
For interpreted languages such as shell, python, R, etc, org-babel allows source blocks to be executed: the code is passed to the interpreter and you have control over what is done with the results of excecution. E.g. place point anywhere in the following block and use C-c C-c to run the code:
Ruby source code
"This file was last evaluated on #{Date.today}"
Results of Ruby evaluation
This file was last evaluated on 2009-08-09
R source code
x = 4
date()
c(5, 10)
Results of R evaluation
5 |
10 |
What happens to the results?
Org-babel provides two fundamentally different modes for capturing the results of code evaluation, specified by the :results header argument:
:results value
This means that the 'result' of code evaluation is defined to be the value of the last statement in the block. Thus with this setting, one can view the code block as a function with a return value. And not only can one view it that way, but you can actually use the return value of one source block as input for another (see later). This setting is the default.
:results output
With this setting, org-babel captures all the text output of the
code block and places it in the org buffer. One can think of this
as a 'scripting' mode: the code block contains a series of
commands, and you get the output of all the commands. Unlike in
the 'functional' mode specified by :results value
, the code
block has no return value. (This mode will be familiar to Sweave
users).
Additional :results settings
Arguments to source code blocks
In addition to evaluation of code blocks, org-babel allows them to be parameterised (i.e. have arguments). Thus source code blocks now have the status of functions.
Inputs for fibonacci-seq
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 18 | 20 |
in the Org-mode buffer this looks like
#+tblname: fibonacci-inputs | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 18 | 20 |
Emacs Lisp source code
(defun fibonacci (n)
(if (or (= n 0) (= n 1))
n
(+ (fibonacci (- n 1)) (fibonacci (- n 2)))))
(mapcar (lambda (row)
(mapcar #'fibonacci row)) fib-inputs)
in the Org-mode buffer this looks like
#+srcname: fibonacci-seq #+begin_src emacs-lisp :var fib-inputs=fibonacci-inputs (defun fibonacci (n) (if (or (= n 0) (= n 1)) n (+ (fibonacci (- n 1)) (fibonacci (- n 2))))) (mapcar (lambda (row) (mapcar #'fibonacci row)) fib-inputs) #+end_src
Results of Emacs Lisp code evaluation
1 | 1 | 2 | 3 | 5 | 8 | 13 | 21 | 34 | 55 |
1 | 3 | 8 | 21 | 55 | 144 | 377 | 987 | 2584 | 6765 |
A meta-programming language for org-mode
Since information can pass freely between source-code blocks and org-mode tables you can mix and match languages using each language for those tasks to which it is suited. This makes Org-mode files with Org-babel into a kind of meta-functional programming language in which functions from many languages can work together.
As an example, lets take some system diagnostics in the shell, and then graph them with R.
- Shell source code
cd ~ && du -sc * |grep -v total
- Results of the shell source code (on my system, grab this org-mode files and try running it on your own)
72 | "Desktop" |
12156104 | "Documents" |
3482440 | "Downloads" |
2901720 | "Library" |
57344 | "Movies" |
16548024 | "Music" |
120 | "News" |
7649472 | "Pictures" |
0 | "Public" |
152224 | "Sites" |
8 | "System" |
56 | "bin" |
3821872 | "mail" |
10605392 | "src" |
1264 | "tools" |
- R source code (which calls the previous shell source code)
pie(dirs[,1], labels = dirs[,2])
- Results of R code
Spreadsheet plugins for org-mode in any language
NOTE: Maybe in-addition-to/in-stead-of this example we should do a more traditional "spreadsheet" example with R [Eric]
Not only can Org-babel pass entire tables of data to source code blocks (see /mirrors/org-mode/src/commit/c23e62cd09049b993b32454837c472fc754f4d49/arguments-to-source-code-blocks), Org-babel can also be used to call source code blocks from within tables using the Org-mode's existing spreadsheet functionality.
In fact the functional test suite for Org-babel is implemented as a
large Org-mode table. To run the entire test suite you simple
evaluate the table C-u C-c C-c
, and all of the tests are run
updating the table with pass/fail statistics.
Here's a sample of our test suite.
functionality | block | arg | expected | results | pass |
---|---|---|---|---|---|
basic evaluation | pass | ||||
emacs lisp | basic-elisp | 2 | 4 | 4 | pass |
shell | basic-shell | 6 | 6 | pass | |
ruby | basic-ruby | org-babel | org-babel | pass | |
python | basic-python | hello world | hello world | pass | |
R | basic-R | 13 | 13 | pass |
code blocks for tests
(* 2 n)
expr 1 + 5
date
"org-babel"
'hello world'
b <- 9
b + 4
Library of Babel
What about those source code blocks which are so useful you want to have them available in every org-mode buffer?
The Library of Babel is an extensible collection of ready-made and easily-shortcut-callable source-code blocks for handling common tasks. Org-babel comes pre-populated with the source-code blocks located in the library-of-babel.org file. It is possible to add source-code blocks from any org-mode file to the library by calling
(org-babel-lob-ingest "path/to/file.org")
Reproducible Research
An article about computational science in a scientific publication is not the scholarship itself, it is merely advertising of the scholarship. The actual scholarship is the complete software development environment and the complete set of instructions which generated the figures.
– D. Donoho
Reproducible Research (RR) is the practice of distributing along with an article of research all data, code, and tools required to reproduce the results discussed in the paper. As such the paper becomes not only a document describing the research but a complete laboratory reproducing the research.
Org-mode already has exceptional support for exporting to html and LaTeX. Org-babel makes Org-mode a tool for RR by activating the data and source code embedded into Org-mode documents making the entire document executable. This makes it not only possible, but natural to distribute research in a format that encourages readers to recreate your results, and perform their own analysis.
Existing RR tools like Sweave provide for the embedding of R code into LaTeX documents. While this is very useful, such documents often still require a large degree of "glue code" in the form of external shell scripts, python scripts, and Makefiles. To my knowledge Org-babl is the only RR tool which allows multiple languages and data to coexist and cooperate inside of a single document.
Literate programming
Let us change our traditional attitude to the con- struction of programs: Instead of imagining that our main task is to instruct a computer what to do, let us concentrate rather on explaining to human beings what we want a computer to do.
The practitioner of literate programming can be regarded as an essayist, whose main concern is with exposition and excellence of style. Such an author, with thesaurus in hand, chooses the names of variables carefully and explains what each variable means. He or she strives for a program that is comprehensible because its concepts have been introduced in an order that is best for human understanding, using a mixture of formal and informal methods that reinforce each other.
– Donald Knuth
Org-babel supports Literate Programming (LP) by allowing the act of programming to take place inside of Org-mode documents. The Org-mode file can then be exported (woven in LP speak) to html or LaTeX for consumption by a human, and the embedded source code can be extracted (tangled in LP speak) into structured source code files for consumption by a computer.
To support these operations Org-babel relies on Org-mode's existing
exporting functionality for weaving of documentation, and on the
org-babel-tangle
function which makes use of Noweb reference syntax
for tangling of code files.
The following example demonstrates the process of tangling in Org-babel.
Literate Programming Example
Tangling functionality is controlled by the tangle
family of
/mirrors/org-mode/src/commit/c23e62cd09049b993b32454837c472fc754f4d49/header-arguments. These arguments can be used to turn tangling on or
off (the default) on the source code block, or the outline heading
level.
The following demonstrates the combination of three source code blocks
into a single source code file using org-babel-tangle
.
The following two blocks will not be tangled by default since they
have no tangle
header arguments.
#+srcname: hello-world-prefix #+begin_src sh :exports none echo "/-----------------------------------------------------------\\" #+end_src
#+srcname: hello-world-postfix #+begin_src sh :exports none echo "\-----------------------------------------------------------/" #+end_src
The third block does have a tangle
header argument indicating the
name of the file to which it should be written. It also has Noweb
style references to the two previous source code blocks which will be
expanded during tangling to include them in the output file as well.
#+srcname: hello-world #+begin_src sh :tangle hello :exports none # <<hello-world-prefix>> echo "| hello world |" # <<hello-world-postfix>> #+end_src
Calling org-babel-tangle
will result in the following being written
to the hello.sh
file.
#!/usr/bin/env sh
# generated by org-babel-tangle
# [[file:~/src/org-babel/org-babel-worg.org::#literate-programming-example][block-16]]
# <<hello-world-prefix>>
echo "/-----------------------------------------------------------\\"
echo "| hello world |"
# <<hello-world-postfix>>
echo "\-----------------------------------------------------------/"
# block-16 ends here
Reference / Documentation
Source Code block syntax
The basic syntax of source-code blocks is as follows:
#+srcname: name #+begin_src language header-arguments body #+end_src
- name
- This name is associated with the source-code block. This is
similar to the
#+TBLNAME
lines which can be used to name tables in org-mode files. By referencing the srcname of a source-code block it is possible to evaluate the block for other places, files, or from inside tables. - language
- The language of the code in the source-code block, valid values must be members of `org-babel-interpreters'.
- header-arguments
- Header arguments control many facets of the input to, evaluation of, and output of source-code blocks. See the Header Arguments section for a complete review of available header arguments.
- body
- The actual source code which will be evaluated. This can be edited with `org-edit-special'.
Header Arguments
- results
-
results arguments specify what should be done with the output of source-code blocks
-
The following options are mutually exclusive, and specify how the results should be collected from the source-code block
- value
- output
-
The following options are mutually exclusive and specify what type of results the code block will return
- vector
- specifies that the results should be interpreted as a multidimensional vector (even if the vector is trivial), and will be inserted into the org-mode file as a table
- scalar
- specifies that the results should be interpreted as a scalar value, and will be inserted into the org-mode file as quoted text
- file
- specifies that the results should be interpreted as the path to a file, and will be inserted into the org-mode file as a link
-
The following options specify how the results should be inserted into the org-mode file
- replace
- the current results replace any previously inserted results from the code block
- silent
- rather than being inserted into the org-mode file the results are echoed into the message bar
-
- exports
-
exports arguments specify what should be included in html or latex exports of the org-mode file
- code
- the body of code is included into the exported file
- results
- the results of evaluating the code is included in the exported file
- both
- both the code and results are included in the exported file
- none
- nothing is included in the exported file
- tangle
-
tangle arguments specify whether or not the source-code block should be included in tangled extraction of source-code files
- yes
- the source-code block is exported to a source-code file named after the basename (name w/o extension) of the org-mode file
- no
- (default) the source-code block is not exported to a source-code file
- other
- any other string passed to the
tangle
header argument is interpreted as a file basename to which the block will be exported
Noweb reference syntax
The Noweb Literate Programming system allows named blocks of code to
be referenced by using a <<code-block-name>>
syntax. When a
document is tangled these references are replaced with the named code.
An example is provided in the /mirrors/org-mode/src/commit/c23e62cd09049b993b32454837c472fc754f4d49/literate-programming-example in this
document.