Automerge fails for two divergent notebook with the same base

[eco32i]

The (minmal) test case:

https://gist.github.com/eco32i/2f9877b8044c28ef8a5ca525342ec73e

With these three notebooks executing the merge command fails:

$ nbmerge test.ipynb test_A.ipynb test_B.ipynb -o test_merged
Traceback (most recent call last):
  File "/home/ilya/.venv/pydata3/bin/nbmerge", line 9, in <module>
    load_entry_point('nbdime', 'console_scripts', 'nbmerge')()
  File "/mnt/data/scipy2016/nbdime/nbdime/nbmergeapp.py", line 109, in main
    r = main_merge(args)
  File "/mnt/data/scipy2016/nbdime/nbdime/nbmergeapp.py", line 43, in main_merge
    m, lc, rc = merge_notebooks(b, l, r, args)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/notebooks.py", line 54, in merge_notebooks
    autoresolve_notebook_conflicts(merged, local_diffs, remote_diffs, args)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/notebooks.py", line 36, in autoresolve_notebook_conflicts
    autoresolve(merged, local_diffs, remote_diffs, strategies, "")
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 346, in autoresolve
    return autoresolve_dicts(merged, local_diff, remote_diff, strategies, path)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 317, in autoresolve_dicts
    newvalue, ldi, rdi = autoresolve(value, le.diff, re.diff, strategies, subpath)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 348, in autoresolve
    return autoresolve_lists(merged, local_diff, remote_diff, strategies, path)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 234, in autoresolve_lists
    newvalue, ldi, rdi = autoresolve(merged[j], le.diff, re.diff, strategies, subpath)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 346, in autoresolve
    return autoresolve_dicts(merged, local_diff, remote_diff, strategies, path)
  File "/mnt/data/scipy2016/nbdime/nbdime/merging/autoresolve.py", line 281, in autoresolve_dicts
    assert set(lcd) == set(rcd)
AssertionError

[martinal]

Thanks a lot for a very useful small test case. I've fixed a couple of issues with the autoresolution based on this. A question for you: what would your expectation be for the merge result of the first cell here? The source code (print(1+3)) is unchanged in A but the transients (execution count and output) change, while the cell is deleted in B. If transients are ignored, a natural result is to delete the cell. If the transients are considered as important data, it must be presented to the user for resolution. We might need options to configure the importance of output changes.

[martinal]

Note that if you run nbmerge now, the output is a valid notebook with the unresolved conflicts stored in the metadata. This is a temporary stage, the mergetool for conflict resolution is work in progress.

[eco32i]

I stumbled upon this while trying to merge two notebooks with real life analyses and then reduced it to what I thought was a minimal test case. From the original notebooks I'd say I want to keep both but this may not be the desired course of action in all cases. May be leaving it up to user to handle is not a bad idea.

[martinal]

This use case currently fails on missing handling of execution count conflict in the mergetool.

[martinal]

Deleting all output here and trying merge results in first cell not checked for deletion in merge gui.