Org-Babel Cheat Sheet

Keyboard Shortcuts

Code Block Shortcuts

Keys Command Effect
C-c C-c org-babel-execute-src-block Execute the code in the current block.
C-c '   Open/close edit-buffer with mode set to match the code-block language.
C-c C-v C-z org-babel-switch-to-session Open a python/ipython console (only works with :session)

Buffer-wide Shortcuts

Keys Command Effect
<s Tab   Create a code block.
C-c C-v C-b org-babel-execute-buffer Execute all code blocks in buffer.
C-c C-v C-f org-babel-tangle-file Tangle all blocks marked to :tangle
C-c C-v C-t org-babel-tangle Seems like an alias for tangle file…

Code Block Headers

This is the subset of headers/header values that I'm interested in right now.

Code to tangle

The pattern I use to tangle (create an external code file) is:

  • python as the language (since I'm not using it with an interactive session, no need for ipython)
  • :noweb tangle is turned on from init.el so that I can substitute code defined elsewhere into the block
  • :tangle <path to file>
  #+begin_src python :tangle literate_python/
    """A docstring for the module"""

    # imports
    import sys

    # constants

    # exception classes

    # interface functions

    # classes


    # internal functions & classes


    if __name__ == "__main__":
	status = main()

Since I have :noweb tangle set, the substitions (e.g. <<literate-main-imports>>) don't get expanded in HTML/Latex output (although they do when you create the python file).

  """A docstring for the module"""

  # imports
  import sys

If you want to show the substitutions when exporting use :noweb yes in the header.

  """A docstring for the module"""

  # imports
  import sys

A named section

The noweb substitution above (<<literate-main-imports>>) worked because there was a named-section (defined here) that it could use:

  #+name: literate-main-imports
  #+begin_src python
    from argparse import ArgumentParser


I now prefer to use :noweb-ref in the header instead of the separate #+name: block.

  #+begin_src python :noweb-ref literate-main-imports
    from argparse import ArgumentParser


The :results header argument declares how to handle what's returned from executing a code block. There are three classes of arguments and you can use up to one of each in the header.

Result Classes

Class Meaning
collection How the results should be collected if there's multiple outputs.
type Declare what type of result the code block will return.
handling How should results be handled.

Collection Class

Option Meaning
value (Default) Uses the value of the last statement in the block (python requires a return statement)
output (:results output) Collects everything sent to stdout in the block.

Type Class

Option Example Meaning
table :results value table Return an org-mode table (vector)
scalar :results value scalar Return exactly the value returned (string)
file :results value file Return an org-mode link to a file
raw :results value raw Return as org-mode command
html :results value html Expect contents for #+begin_html
latex :results value latex Expect contents for #+begin_latex
code :results value code Expect contents for #+begin_src
pp :results value pp Expect code and pretty-print it

Handling Class

Option Example Meaning
silent :results output silent Don't output in org-mode buffer
replace :results output replace (Default) Overwrite any previous result
append :results output append Append output after any previous output
prepend :results output prepend Put output above any previous output


This argument tells org-babel what to put in any exported HTML or Latex files.

Option Example Meaning
code :exports code (default) The code in the block will be included in the export.
results :exports results The result of evaluating the code will be included.
both :exports both Include code and results in the file.
none :exports none Don't include anything in the file.

Running Tests

Say there was another section in the document that tangled a test-file (named to test our main source file. Once both are tangled you can run it in the document using sh as the language. The org-mode documentation shows a more complex version of this which builds a pass-fail table, but that's beyond me right now.

   #+name: shell-run-pytest
   #+begin_src sh :results output :exports both
   py.test -v literate_python/
============================= test session starts ==============================
platform linux -- Python 3.5.1+, pytest-3.0.5, py-1.4.32, pluggy-0.4.0 -- /home/cronos/.virtualenvs/nikola/bin/python3
cachedir: .cache
rootdir: /home/cronos/projects/nikola/posts, inifile: 
plugins: faker-2.0.0, bdd-2.18.1
collecting ... collected 1 items

literate_python/ PASSED

=========================== 1 passed in 0.06 seconds ===========================

Specific Block Cases

Plant UML

Besides setting the language to plantuml you need to specify and output-file path and set :exports results so that the actual plantuml code won't be in the exported document but the diagram will.

#+begin_src plantuml :file literate_python/literateclass.png :exports results
skinparam monochrome true

LiterateClass : String who
LiterateClass : String ()


The main thing to remember for ob-ipython is that you need to run it as a :session. I didn't do it for most of the examples, but I've found since I first wrote this that using named sessions makes it a lot easier to work. Otherwise you might have more than one buffer with an org-babel document and they will be sharing the same ipython process, which can cause mysterious errors.

#+begin_src ipython :session session1
  # python standard library
  import os

When using pandas most of the methods produce values, but the info method instead prints to stdout so you have to specify this as the :results or it will popup a separate buffer with the output.

#+begin_src ipython :session session1 :results output

When you create figures, besides making sure that you use the %matplotlib inline magic, you also need to specify a file path where matplotlib can save the image.

#+BEGIN_SRC ipython :session session1 :file "images/ocean_proximity_count.png"
figure = seaborn.countplot(x="ocean_proximity", data=housing)

Set Up


I'm using ob-ipython to use jupyter/ipython with org-babel so you have to install it (I used MELPA). In addition you need to install the python dependencies, the main ones being ipython and jupyter. Additionally, I use elpy (also from MELPA) which has its own dependencies. I think the easiest way to check and see what elpy dependencies you need is to install elpy (there's two components, an emacs one you install from melpa and a python component you install from pip) then run M-x elpy-config to see what's missing.


Since I mentioned ob-ipython and elpy I'll list what I have in my init.el file for elpy and org-babel.


;; elpy
(setq elpy-rpc-backend "jedi")
(eval-after-load "python"
 '(define-key python-mode-map "\C-cx" 'jedi-direx:pop-to-buffer))


;; org-babel
;;; syntax-highlighting/editing
(add-to-list 'org-src-lang-modes '("rst" . "rst"))
(add-to-list 'org-src-lang-modes '("feature" . "feature"))

;;; languages to execute/edit
 '((ipython . t)
   (plantuml . t)
   (shell . t)
   (org . t)
   ;; other languages..

;;; noweb expansion only when you tangle
(setq org-babel-default-header-args
      (cons '(:noweb . "tangle")
	    (assq-delete-all :noweb org-babel-default-header-args))

;;; Plant UML diagrams
(setq org-plantuml-jar-path (expand-file-name "/usr/share/plantuml/plantuml.jar"))

;;; execute block evaluation without confirmation
(setq org-confirm-babel-evaluate nil)   

;;; display/update images in the buffer after evaluation
(add-hook 'org-babel-after-execute-hook 'org-display-inline-images 'append)

Integrating with Nikola/Sphinx