Notebook Cherry Picker Extension

Description

Simple Notebook & nbconvert extension demonstrating how to cherry pick cells from a notebook.

• Old video demonstration of notebook GUI: http://youtu.be/iACSt4_vv9Y
• Old video demonstration of cherry picking operation: http://youtu.be/7d-2IWZ_vLs

Installation

1. If you don't have a user profile already, go ahead and create one by running ipython profile create. You can then figure out where your profile is located by running ipython profile list. It will be listed under the users section. Mine is ~/.config/ipython/profile_default/. Advanced users may not want to override custom.js, that's fine as long as a reference to cell_tags is added to the existing custom.js
2. Put the files in /static/custom/ into the same subdir in your profile directory. You may need to create it if it doesn't already exist. For example, on my computer, the files reside in ~/.config/ipython/profile_default/static/custom/.
3. Refresh any notebooks you have open, the cell toolbar should now be available in drop-down menu on the main toolbar. Also, a new tag button should be available on the main toolbar.

Usage

Part 1 - Notebook

This plugin works by allowing the user to tag each cell in the notebook. Tags are specified on a cell to cell basis via the new tags cell toolbar. The information is stored in the cell metadata, and can be viewed using the Raw Edit button on the default toolbar.

Part 2 - nbconvert

When all of the cells are tagged as desired, one must run nbconvert with a special writer and preprocessor (included) to cherry-pick the desired cells.

• Make sure that the cherry_picking_preprocessor.py and notebook_copy_writer.py are accessible via your PYTHONPATH. The easiest way to do this is to place them in the cwd (usally the same directory as the notebooks that you want to cherry pick from).
• Make sure that the ipython_nbconvert_config.py file (included) is placed in the cwd. This is a configuration file who's only purpose is to shorten the CLI. If you want

Call nbconvert, specify the names of the notebooks you want to convert (glob supported) and the tags you want to pick out of the document.

ipython nbconvert example.ipynb --CherryPickingPreprocessor.expression="instructor or exercise"

where

• example.ipynb would be replaced by a space separeted list of notebook names
• instructor or exercise would be replaced by the boolean expression used to select cells.

Example:

ipython nbconvert notebook*.ipynb --CherryPickingPreprocessor.expression="(manager or supervisor) and secret"

If you want to shorten the command, you can add the expression to the local config file. You can also have multiple config files, each one with separate tags. You can then load the config file you want with --config myConfig.py at the commandline.