Pondering generating notebooks from nbdime again (as
per https://github.com/jupyter/nbdime/issues/251 ) I was wondering if
there's a simple way of getting numeric cell key values from a notebook?
eg if ai have a setup
fn1="test1.ipynb"
fn2="test2.ipynb"
a = nbformat.read(fn1, as_version=4)
b= nbformat.read(fn2, as_version=4)
I seem to be able to get a list of diffs
as: nbdime.diffing.diff_notebooks(a, b)
nbdime.diffing.generic.diff_dicts(a, b)
though I don't really understand the datastructure (eg multiple levels of
"diff" inside a "diff" per things like: {'diff': [{'diff': [{'diff':
[{'diff': [{'diff': [{'diff': [{'key': 46, }
I assume that the 'key' uniquely identifies a line in the notebook a - if
so, is there anything that takes notebook and returns it as a data
structure with 'key' elements identified according to the same numbering
scheme? Then I could print out the (unchanged) elements from a if the key
is not in the diff file, or print the diff element if the key is in the
diff file?
eg as a roundabout case in point - I can print notebook a as:
nbdime.prettyprint.pretty_print_notebook(a)
and diff between notebook a and b:
nbdime.prettyprint.pretty_print_notebook_diff(fn1,fn2,a,d)
that gives a display akin to nbdiff-web with "Hide unchanged cells" checked:
But how would I go about doing a pretty print to display the equivalent of
nbdiff-web
with "Hide unchanged cells" unchecked (i.e. displaying unchanged cells too)?
--tony
PS context is:
What I have in mind exploring is a set of functions equivalent to
pretty_print_diff_entry()
etc along the lines of nb_create_diff_entry() that add metadata annotated
cells to an output notebook rather than as colour coded lines in an output
text stream. Cell metadata state would identifying the state of the cell
(added, removed, etc). I'd also like a switch to be able to add
(optionally) unchanged cells to the output notebook, or just unchanged
notebook cells with a particular metadata element set in the original
notebook (so eg I could force a cell to always appear in the output diff
whether it was changed or not).
My use case is trying to explore the space of instructor marking of student
notebooks where students are provided with a template notebook which may
contain lots of text. Idea would be to create an diff-notebook that
contains just student modified cells and some heading cells from the
original notebook. (Thinks: actually, I could do that anyway - add a
'stripme' metadata element to cells in notebook supplied to students then
run student returned notebooks through a processor that creates a child
notebook without the 'stripme' cells.)
The nbgrader doesn't quite work for my setting (and I'm not sure what
workflow would). Something around diffing feels right, but I want to try to
keep things as diff-revealing jupyter notebooks rather than a diff
revealing text files.
I think I need to walk the dog again to try to clarify my thoughts a bit
more!
--
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 [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jupyter/5e895be4-6ad6-4fec-a751-dbede9d68b30%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
