IPython extension to mask sensitive data

Simple ipython extension to mask sensitive data like credentials from notebook cell outputs. This extensions may be useful to avoid inadvertently displaying sensitive data in public notebooks. The extension modifies the ipython display system to try and mask any occurence of any of the given strings that appear in textual outputs from printing, logging and the native display system.

Usage

To load the extension use the magic command %load_ext nbmask. The extension will automatically mask the notebook display textual outputs without any further magic commands. When first loaded the extension will mask just the username.

import os
from pathlib import Path

%load_ext nbmask

username = os.getenv('USER')
print("My name is {username}!")
# >>> My name is ...!

documents = Path(f"~/Documents").expanduser()
documents
# >>> PosixPath('/Users/.../Documents')

You can add more secrets with the %nbmask magic line command and a string parameter using python builtin variable expansion syntax.

TOKEN = my_secret_token()

%nbmask "$TOKEN"

credentials = dict(user=username, token=TOKEN)

credentials
# >>> {'user': '...', 'token': '...'}

The extension modifies standard output, standard error and also the default logging handlers.

import logging

logging.basicConfig(level="DEBUG")

logging.debug("Token is %s", TOKEN)
# >>> DEBUG:root:Token is ...

Example

See nbmask-tests.ipynb in extras

Installation

You can install the current version of this package with pip

pip install nbmask

Changelog

0.05

• Now using uv_build backend
• Added testing with pytest and nbmake
• Requires python >= 3.10

0.0.4

• Cell magic %%masked is no longer needed. Will be removed

0.0.3

• Masking pattern is cached