A python iterator through git commits.

Current development state: very early

Just started to make a reusable module from the source that was previously copied from one project to the next.

pip install git+https://github.com/defgsus/giterator

from giterator import Giterator

repo = Giterator(PATH_TO_REPO)

# loop through all commits in historical order
for commit in repo.iter_commits():
    # all infos attached
    print(commit.committer_date, commit.hash)
    # loop through all files that have changed in this commit
    for file in commit.iter_commit_files():
        print(file.name, file.content)

The Giterator.iter_commits() method returns a generator which internally uses the git log command. While the git stdout stream is read, new Commit objects are yielded one-by-one. So iterating through millions of commits does not generally lead to memory problems.

The Giterator.iter_files() method, which is the base for all other iter_files methods, currently needs to hold all files in memory. Maybe this can be optimized.