FOI: use libgit/pygit2 for extracting "core" metadata?

[yarikoptic]

ATM relies on git log -1 per file to extract information about each file. I wonder if it might be more efficient to implement it through "parsing" git log --stats thus going through the history until all files in the tree are covered. To avoid parsing of git log I thought that may be libgit could be used.

here is ipython history where I investigated how to go through history of commits and get to "stats" per each commit:

In [26]: history
import rich
import pygit2
repo = pygit2.Repository('.')
c = next(repo.walk(repo.head.target))
rich.inspect(c)
pygit2.diff(c.hex)
repo.diff(c.hex)
diffstat = repo.diff(c.hex)
rich.inspect(diffstat)
rich.inspect(diffstat.stats)
rich.inspect(diffstat.deltas)
rich.inspect(next(diffstat.deltas))
c.hex
rich.inspect(c)
diffstat = repo.diff(c.hex, c.parents[0].hex)
rich.inspect(next(diffstat.deltas))
rich.inspect(next(diffstat.stats))
rich.inspect(next(diffstat))
rich.inspect(diffstat)
rich.inspect(diffstat.stats)
rich.inspect(next(diffstat.stats))
rich.inspect(next(diffstat.deltas))
rich.inspect(next(diffstat.deltas).new_file)
history