In [1]:
%load_ext load_style
%load_style talk.css
I recommend that you install the latest development version of runipy from the github repository with the following:
pip install https://github.com/paulgb/runipy/archive/master.zipAnother solution (for IPython version 3) is to use nbconvert, which now comes with an option to run a notebook in place and save the outputs in the same notebook, the (rather verbose) syntax is:
ᐅ ipython nbconvert notebook_to_execute.ipynb --ExecutePreprocessor.enabled=True --to notebook --output notebook_to_execute.ipynbHowever it does not seem to handle passing environment variables as runipy does ...
see http://ipython.org/ipython-doc/dev/whatsnew/version3.html for the list of changes in the version 3.0 of IPython and the IPython notebook, see in particular the backwards incompatible changes section.
The following comes from the doc of Runipy and goes through its major use cases
The IPython notebook provides an interactive interface to a Python interpreter.
Literate programming: the IPython notebook is an ideal format for
writing "literate" programs, in which the code is part of a larger
multi-media document. runipy lets you run such programs directly,
without first converting to a pure Python script.
Report generation: runipy can run the notebook and convert it
into HTML in one go, making it an easy way to automate reports.
Data pipeline: if you use IPython notebooks to create a data
pipeline, runipy lets you automate that pipeline without losing
the notebook formatting.
runipy currently supports IPython versions 2.3.x and the current
development version of 3.x.
The easiest way to install runipy is with pip:
$ pip install runipy
To run a .ipynb file as a script, run:
$ runipy MyNotebook.ipynb
To save the output of each cell back to the notebook file, run:
$ runipy -o MyNotebook.ipynb
NOTE: The notebook seems to be automatically converted back to version 3 if it was in version 4.
To save the notebook output as a new notebook, run:
$ runipy MyNotebook.ipynb OutputNotebook.ipynb
To run a .ipynb file and generate an HTML report, run:
$ runipy MyNotebook.ipynb --html report.html
You can pass arguments to the notebook through environment variables. The use of environment variables is OS- and shell- dependent, but in a typical UNIX-like environment they can be passed on the command line before the program name:
$ myvar=value runipy MyNotebook.ipynb
Then in the notebook, to access myvar:
from os import environ
myvar = environ['myvar']
environ is just a dict, so you can use .get() to fall back
on a default value:
from os import environ
myvar = environ.get('myvar', 'default!')
runipy can read stdin and stdout and sit in a UNIX pipeline:
::
$ runipy --stdout < MyNotebook.ipynb > OutputNotebook.ipynb
$ cat MyNotebook.ipynb | runipy --stdout > OutputNotebook.ipynb
It is also possible to run IPython notebooks from Python, using:
from runipy.notebook_runner import NotebookRunner
from IPython.nbformat.current import read
notebook = read(open("MyNotebook.ipynb"), 'json')
r = NotebookRunner(notebook)
r.run_notebook()
and you can enable pylab with:
r = NotebookRunner(notebook, pylab=True)
The notebook is stored in the object and can be saved using:
from IPython.nbformat.current import write
write(r.nb, open("MyOtherNotebook.ipynb", 'w'), 'json')
run_notebook() takes two optional arguments. The first, skip_exceptions, takes a boolean value (False by default). If True, exceptions will be ignored and the notebook will continue to execute cells after encountering an exception. The second argument is progress_callback, which must be either None or a function that takes one argument. This function is called after execution of each cell with the 0-based index of the cell just evaluated. This can be useful for tracking progress of long-running notebooks