The latest version of nbformat>=5.7.0 emits a message when a notebook has no cell ids.
I'd like to see what we could do about that. Maybe I'd prefer not to include the cell ids in the text notebooks by default, but we could still give an option to do so.
Also, pandoc>=2.7 does support cell ids, but Jupytext currently filters out the cell ids when saving to a text format , we should think about that.
Other possibilities would be
generate simple ids (like: unnamed_code_cell_1) for cells that don't have one, and ignore the ids that matches the pattern when we write the text version of the notebook
save the non trivial ids as the 'title' of the cell
The latest version of
nbformat>=5.7.0emits a message when a notebook has no cell ids.I'd like to see what we could do about that. Maybe I'd prefer not to include the cell ids in the text notebooks by default, but we could still give an option to do so.
Also,
pandoc>=2.7does support cell ids, but Jupytext currently filters out the cell ids when saving to a text format , we should think about that.Other possibilities would be
unnamed_code_cell_1) for cells that don't have one, and ignore the ids that matches the pattern when we write the text version of the notebook