nbcomet is a Jupyter Notebook extension that tracks Notebook history. It consists of both a server side extension (.py) and a client-side nbextenion (.js). Both must be installed for the tracking to work properly.
For your convenience, nbcomet's server and nbextenion are contained in a single python package that can be installed with pip or conda. Simply run one of the following commands in your terminal:
pip install nbcomet
# or install with conda
conda install -c adamrule nbcomet
Note: If you have multiple versions of python, make sure you download
nbcomet onto the version of python you use to run your notebook server. You can
check which version of python is running your notebook server by selecting the
"Help > About" menu item in any running notebook. If this version is different
from your default python version (e.g. what shows up when you run python -v in
your Terminal), you can specify which version of python pip installs the the
extension onto by modifying the pip command:
pip2.7 install nbcomet
pip3.5 install nbcomet
For more detailed assistance with loading Jupyter Notebook extensions, see the Jupyter Extension documentation.
While nbcomet is now installed, you need to tell Jupyter to run it every time you run a notebook. You can use the following three commands to finish installing and enabling the notebook and server extensions:
jupyter nbextension install --py nbcomet
jupyter nbextension enable --py nbcomet
jupyter serverextension enable --py nbcomet
You can check that nbcomet installed correctly by running:
jupyter serverextension list
jupyter nbextension list
By default, Comet with store its data in ~/.jupyter/nbcomet; You can change
this folder by editing the notebook.json configuration file in your
~/.jupyter/nbconfig folder to include a line specifying your data directory.
For example: "Comet": {"data_directory": "/full/path/to/directory" },.
Comet tracks how your notebook changes over time. It does so by:
- tracking the occurrence of actions such as creating, deleting, moving, or executing cells
- tracking how your notebook changes as a result of these actions
Comet tracks this information in two ways:
- periodically saving a full version of the notebook
- saving the name and time of every action performed in the notebook to an sqlite database
Comet is a research tool designed to help scientists in human-computer
interaction better understand how people use Jupyter Notebooks. It is primarily
a recording tool with limited support for visualizing or reviewing the recorded
data. You can see a basic visualization of your notebook activity by selecting
Comet > See Comet Data in your Jupyter Notebook menu.
