a nose plugin for IPython notebooks
Project description
a nose plugin for finding and running IPython 2/3 notebooks as nose tests.
What it can’t do in terms of setup and tearDown, nosebook makes up for in simplicity: there is no %%nose magic, no metadata required: the notebook on disk is the “gold master”.
This makes it ideal for decreasing the burden of keeping documentation up to date with tests by making a single set of notebooks into both rich, multi-format documentation and a simple part of your test suite.
How does it work?
Each notebook found according to `nosebook-match <#nosebook-match>`__ is started with a fresh kernel, based on the kernel specified in the notebook. If the kernel is not installed, no tests will be run and the error will be logged.
Each code cell that matches `nosebook-match-cell <#nosebook-match-cell>`__ will be executed against the kernel in the order in which it appears in the notebook: other cells e.g. markdown, raw, are ignored.
The number and content of outputs has to match exactly, with the following parts of each output stripped:
execution/prompt numbers, i.e. [1]:
tracebacks
Non-deterministic output, such as with _repr_ methods that include the memory location of the instance, will obviously not match every time. You can use `nosebook-scrub <#nosebook-scrub>`__ to rewrite or remove offending content.
Configuring nosetests to use nosebook
These options can be specified in your nose config file, or as long-form command line arguments, i.e. --with-nosebook.
with-nosebook
nosetests will look for notebooks that seem like tests, as configured with `nosebook-match <#nosebook-match>`__.
Default: False
# Basic usage
!nosetests --with-nosebooknosebook-match
A regular expression that tells nosebook what should be a testable notebook.
Default: ``.*[Tt]est..ipynb$``*
# Run against all notebooks... probably not a good idea, but maybe a great idea
!nosetests --with-nosebook --nosebook-match .*.ipynbnosebook-match-cell
A regular expression that will be replaced throughout the expected outputs and generated outputs.
Default: None
# will run cells where tests are defined... but you should probably run them, too
!nosetests --with-nosebook --nosebook-match .*Simple.* --nosebook-match-cell '(def|class).*[Tt]est'nosebook-scrub
A regular expression that will be replaced throughout the expected outputs and generated outputs.
Default: None
# you can't fail if you don't try
!nosetests --with-nosebook --nosebook-scrub .+For multiple scrub values, you can pass a JSON-formatted list of regular expressions or object of pattern-replacement pairs that will be replaced. When passed in via the command line, you’ll have to escape special characters: using a .noserc config file makes this easier.
# there are only 10 kinds of tests...
!nosetests --with-nosebook --nosebook-scrub='["0", "1"]'# 0 is equally good
!nosetests --with-nosebook --nosebook-scrub='{"\\d+": "0"}'Contributing
Issues and pull requests welcome!
License
nosebook is released as free software under the BSD 3-Clause license.
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for nosebook-0.4.0-py2.py3-none-any.whl
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 0dae949e62670425bbc477643afa6a80334521d90e8471b6eeced1c92dc35625 |
|
| MD5 | f153313f5fddd9f0a4f9e5e4eb41287d |
|
| BLAKE2b-256 | d578e45864e6a61ff83ecfa5b772a8b8836683a0a326da6cfc887b2c62434619 |
