runnb

A lean utility to run the jupyter notebooks from commandline with stdout outputs


Keywords
jupyter, notebook
License
MIT
Install
pip install runnb==1.0.2

Documentation

runnb - The run-notebook script

runnb is a script to run jupyter notebooks from the command-line. It is a nbconvert API wrapper. Notes on executing notebooks can be found here. You may also want to see the format specification of jupyter notebooks nbformat.

Current version of runnb is implemented and tested under python 2, 3 and jupyter 4.3.0.

Installation

To install pip package,

pip install runnb

or to install latest version,

pip install --upgrade git+https://github.com/InnovArul/runnb

Usage

runnb [options] <path/to/notebook.ipynb>

Options

  • -h --help Display help message.
  • -e --allow-error Allow error during single cell and continue running.
  • -n --no-stdio Don't recover STDIO to command line. (You may not see printed messages.)
  • -a --allow-not-trusted Run the notebook even not trusted.
  • -t --to=<path/to/notebook.out.ipynb> Save the executed notebook to a specific file.

Examples

Assume we have a notebook test.ipynb. Simply run the notebook from command-line:

runnb test.ipynb

If we want to run the notebook not being break by possible error in some cell, pass the --allow-error flag:

runnb --allow-error test.ipynb

Or do it with the shortcut -e:

runnb -e test.ipynb

If we wanted to export the executed notebook, use --to:

runnb --to=test.out.ipynb test.ipynb

or using its shortcut -t:

runnb --ttest.out.ipynb test.ipynb

By default, we recover the output of notebook to the command-line by boardcasting the input/output/error stream. This is done by adding a small snippet of code handling sys.stdin, sys.stdout, and sys.stderr. A tiny Tee Class is used to support such behavior. Make sure tee.py is in our python search path. If doing this is not preferable in certain scenario, we can turn it off by passing --no-stdio flag:

runnb --no-stdio test.ipynb

or using its shortcut -n:

runnb -n test.ipynb

However, this could mean we may not see output from the execution on the command-line. It could be a good idea to save those outputs to a output notebook by using --to flag:

runnb --no-stdio --to test.out.ipynb test.ipynb

The most (no exaggeration) dangerous thing we can try here is to run a not trusted notebook. By default, running a not-trusted notebook leads to a DeprecationWarning that stop it from running. However, if one insists to, we can pass --allow-not-trusted flag to allow running.

runnb --allow-not-trusted not_trusted.ipynb

Deprecation message will be displayed but it would not block running. The executed not-trusted notebook could also be saved by using --to. We will unsign the outcome of not-trusted notebook and mark it not trusted. However, the best practice is do not run not trusted notebooks. At least we should review the notebook first, and sign it trusted if it is safe. Make sure we understand what is happenning and what we are doing.

Lib usage

runnb(nb_path, allow_errors=False, no_stdio=False, to_file=None)

Run a notebook from current path.

Parameters:

  • nb_path (str) - path to a notebook.
  • allow_errors (bool) - Wheither to allow error during single cell and continue running. When set to True, the notebook will continue running following cells if error presents in some cell. Default False.
  • no_stdio (bool) - Wheither to stop recovering STDIO to default. When set to True, default STDIO (usually command-line) will not be recovered. Results can be found only in the output notebook. Default False.
  • to_file (str) - Path to which file the executed notebook will be saved to. The notebook will not be save if set to None or empty string. Default None.

Raises:

  • DeprecationWarning - Raise if the notebook is not trusted and the warning is not filtered with warnings module.