CLI to compare 2 kastore files

[hyanwong]

A few times now I have wanted to work out why two almost identical tree sequences have differing file sizes. The most comprehensive way to do this is python3 -m kastore ls -l file.kastore on both files, followed by comparing the output. It would be lovely to have a command-line option that took 2 files and showed where the differences lay, perhaps showing a 3 column layout with file1, file2, difference, where arrays with the same name in file1 and file2 have the difference shown, and arrays only present in one of the other file are listed as "only in file1" or "only in file 2". Colouring the largest value in any row in red would be the cherry on the cake. None of this is essential, though.

[jeromekelleher]

Writing a general diff is not a trivial thing to do @hyanwong - do you have some tools from the standard library in mind for implementing this?

[grahamgower]

Looking at what unittest does in this case might be useful.
https://github.com/python/cpython/blob/3c4fc864ce931db90214c54d742062a80dbef7c4/Lib/unittest/case.py#L922
It essentially amounts to truncating this output:

            difflib.ndiff(pprint.pformat(seq1).splitlines(),
                          pprint.pformat(seq2).splitlines()))