Easy to use test framework for Jupyter Notebooks.
- Runs notebook top to bottom and flags execution errors if any
- Runs unittest present in your notebook code cells
- Runs doctest present in your notebook code cells
- Start testing notebooks without writing a single line of test code
- Multithreaded execution for quickly testing a set of notebooks
- Executes every Notebook in a fresh kernel to avoid hidden state problems
- Primarily a command line tool that can be used easily in any Continuous Integration (CI) system
- Soon to be part of ReviewNB's CI system that automatically runs treon everytime you push notebook changes to GitHub
pip install treon
$ treon
Executing treon version 0.1.0
Recursively scanning /workspace/treon/tmp/docs/site/ru/guide for Notebooks...
-----------------------------------------------------------------------
Collected following Notebooks for testing
-----------------------------------------------------------------------
/workspace/treon/tmp/docs/site/ru/guide/keras.ipynb
/workspace/treon/tmp/docs/site/ru/guide/eager.ipynb
-----------------------------------------------------------------------
Triggered test for /workspace/treon/tmp/docs/site/ru/guide/keras.ipynb
Triggered test for /workspace/treon/tmp/docs/site/ru/guide/eager.ipynb
test_sum (__main__.TestNotebook) ...
ok
test_sum (__main__.TestNotebook2) ...
ok
test_sum (__main__.TestNotebook3) ...
ok
----------------------------------------------------------------------
Ran 3 tests in 0.004s
OK
-----------------------------------------------------------------------
TEST RESULT
-----------------------------------------------------------------------
/workspace/treon/tmp/docs/site/ru/guide/keras.ipynb -- PASSED
/workspace/treon/tmp/docs/site/ru/guide/eager.ipynb -- PASSED
-----------------------------------------------------------------------
2 succeeded, 0 failed, out of 2 notebooks tested.
-----------------------------------------------------------------------
Usage:
treon
treon [PATH] [--threads=<number>] [-v]
Arguments:
PATH File or directory path to find notebooks to test. Searches recursively for directory paths. [default: current working directory]
Options:
--threads=<number> Number of parallel threads. Each thread processes one notebook file at a time. [default: 10]
-v --verbose Print detailed output for debugging.
-h --help Show this screen.
--version Show version.
You just need to add tests as shown below & treon would execute them and report the result on the console. See this for more details on how to write unittest.
You just need to add tests as shown below & treon would execute them and report the result on the console. See this for more details on how to write doctest.
- You need to run treon from environment (virtualenv/pipenv etc.) that has all the dependcies required for Notebooks under test
- treon only works with python3+ environments and uses python3 kernel for executing notebooks
For development, you may use below to create a Python interpreter that resides in venv in the current working directory, and to install all of treon's dependencies:
$ virtualenv venv
$ source venv/bin/activate
$ pip install -e .
$ pip install -r requirements-dev.txt
$ treon --help # should work
Because the script installs the package as editable, you can make changes in the source tree and use the treon command to immediately validate them. If this does not appear to work, check that you are using a the proper virtual environment, and that the package is indeed installed in editable mode:
$ which treon # should point into your virtualenv
/path/to/my/venv/bin/treon
$ pip list --local | grep treon # should point to the source tree
treon 0.1.2 /workspace/treon
Please refer to the Makefile for supplementary development tasks.
In particular, the following targets may be relevant when validating changes before committing:
$ make lint # check treon's source for code style errors
$ make test # run all tests
Our aim at ReviewNB is to make notebooks a first class entity in the production workflow. We've built a code review system for Notebooks. The next step is to build a CI pipeline & treon is the core tool in that effort. It is licensed librerally (MIT) & I foresee it being used as an independent tool as well. You can use it locally and/or integrate with CI system of your choice.
For motivation, checkout Netflix's blog to see how notebooks are graduating from scratchpad to a part of production workflow.
If you see any problem, open an issue or send a pull request. You can write to team@reviewnb.com for any questions.