[ANN] nbformat 4.4.0

[M Pacer]

We are pleased to announced the release of nbformat 4.4.0. 

On top of the usual doc improvements and bug fixes, this minor release does a few things especially worthy of note by the wider Jupyter community.

It introduces in cell metadata the metadata.jupyter namespace for jupyter official metadata fields, and includes the source_hidden and outputs_hidden fields for front-ends to store whether input or output is hidden when viewing a notebook. These fields will not be used to hide input or outputs on export using nbconvert.

The nbformat.validate function now has a new optional argument relax_add_props, which when set to true will allow validating a notebook that has properties in addition to those required by nbformat. This is particularly useful for the creation of intermediate representations in applications such as nbconvert.

You can now assign dict-like values to attributes on nbformat.NotebookNode objects and they will be converted to be NotebookNode objects. This also works if you update a NotebookNode object with a dict-like object (or keys with dict-like values).

Please see the change-log for more information: http://nbformat.readthedocs.io/en/latest/changelog.html#id1

The conda-forge update is underway (https://github.com/conda-forge/nbformat-feedstock/pull/8).

We thank the following 7 authors who contributed 44 commits to this release.

* Benjamin Ragan-Kelley

* Brian E. Granger

* Jason Grout

* Kyle Kelley

* M Pacer

* madhu

* Thomas Kluyver

Cheers,

The nbformat team

[Brian Granger]

Hmm, I think there has been a significant miscommunication about the
purpose of the `source_hidden` and `outputs_hidden` metadata.

Here is a summary of my understanding:

https://github.com/jupyterlab/jupyterlab/issues/2740

I remember us specifically talking as a groupt at the in person
meeting and it being said that the entire purpose of persisting this
metadata was so that nbconvert and other tools could work with it.
nbconvert may have different ways of hiding content based on tags, but
I don't see why it wouldn't respect this notion.

More importantly, metadata state that is stored in the document can't
be used as the view state as it should not be shared between users in
the real-time collab setting. My comments in the above issue address
these subtleties. Most importantly, the PR that implemented these
flags does not mention the restrictions about these not being used in
nbconvert (https://github.com/jupyter/nbformat/pull/94/files).

I realize you have been asking for feedback on how nbconvert handles
these things and I apologize my time has been so sparse.

> --
> You received this message because you are subscribed to the Google Groups
> "Project Jupyter" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to jupyter+u...@googlegroups.com.
> To post to this group, send email to jup...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/jupyter/CAM3SX44LyRYfuuyoSvbYEqbUcenC-0mfox5EW3GzJ%2B8LMPKo3w%40mail.gmail.com.
> For more options, visit https://groups.google.com/d/optout.



--
Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgra...@calpoly.edu and elli...@gmail.com