Skip to content

Version control, test, and CI/CD your Python Jupyter notebooks.

License

Notifications You must be signed in to change notification settings

DataBiosphere/herzog

Repository files navigation

herzog

Write your Python Jupyter notebooks in herzog!

herzog scripts are pure Python. This means version control, pull requests, CI/CD, and more.

Great! How does it work?

Usage

Use context managers to define Jupyter cells:

with herzog.Cell("python"):
    print("Hello herzog")

Want to tell your users what to do? Make a markdown cell:

with herzog.Cell("markdown"):
    """
    # How cool is this notebook?
    So cool!
    """

Use the herzog CLI to generate a notebook from your Python script:

herzog path/to/my/cool_script.py > path/to/my/cool_notebook.ipynb

Example

An example herzog script is shown below, along with the generated notebook. Both the herzog source and generated notebook can be found in the herzog repository.

Note that everything outside herzog context managers is not included in the notebook. That's where you put tests, developer notes, salacious accusations, or anything else.

import herzog

with herzog.Cell("markdown"):
    """
    # Fibonacci:
    Recursively generate terms from the Fibonacci sequence.
    """

with herzog.Cell("python"):
    def fibonacci(term: int) -> int:
        """
        0-indexed fibonacci
        """
        assert 0 <= term
        if 1 < term:
            return fibonacci(term - 1) + fibonacci(term - 2)
        else:
            return term

# Code outside herzog context does not appear in generated notebooks.
try:
    fibonacci(-1)
except AssertionError:
    # expected
    pass
expected_fibonacci = (0, 1, 1, 2, 3, 5, 8, 13, 21)
for i in range(1, len(expected_fibonacci)):
    assert expected_fibonacci[i] == fibonacci(i)

golden_ratio = (1 + 5 ** 0.5) / 2
assert 1e-10 > abs(golden_ratio - fibonacci(26) / fibonacci(25))

Rendered Fibonacci notebook

Installation

pip install herzog

Links

Project home page GitHub
Package distribution PyPI

Bugs

Please report bugs, issues, feature requests, etc. on GitHub.

About

Version control, test, and CI/CD your Python Jupyter notebooks.

Resources

License

Stars

Watchers

Forks

Packages

No packages published