Python: Finding a mapping of all commits to their diffs for a given git repository

Question

I am creating a tool for analysis of a Git repository, but have stumbled at something that should (seemingly) be quite simple.

I want to create a mapping of commits to diffs (i.e. actual blob changes line by line for the given commit); I have tried using GitPython but haven't had any success. I would like to achieve something like this:

def get_all_commits(REPO_URL):
 chromium_repo = Repo(REPO_URL)
 commits = list(chromium_repo.iter_commits())
 commmit_diffs = {}
 for commit in commits:
 diff = # get all blob changes for commit
 commit_diffs[commit.hexsha] = diff
 return commit_diffs

but am not sure how to get all blob changes for a given commit. commit_diffs would be in the form:

{ 232d8f39bedc0fb64d15eed4f46d6202c75066b6 : '<String detailing all blob changes for given commit>' }

Any help would be great.

• Lex Scarisbrick: It would be helpful if you provided some sample code with output and describe how the output should be different. It looks there is a way to get diff information, but perhaps it doesn't provide what you need?
• Fraser Price: @LexScarisbrick This is the first time I've used this library so I'm really not sure where to start; I've added a code sample to give some sort of idea
• Nils Werner: Be aware that this code will probably consume all your RAM as the chromium repo is quite big. That being said, have you tried commit.diff()?
• Fraser Price: @NilsWerner Yes I'm aware of memory issues; is a one time computation which will be stored in a db. Yes, although is a diff object not just a hash of the entire tree? How would I find differences between both trees efficiently? As you mentioned, the chromium repo is huge; having to compare the entire tree for every commit seems pretty ludicrous, unless there is something I'm missing

Answer 1

I was unaware of the git diff <commit_a> <commit_b> command. The following (I think!) solves the issue:

def get_all_commits(REPO_URL):
 repo = Repo(REPO_URL)
 commits = list(repo.iter_commits())
 commmit_diffs = {}
 for index, commit in enumerate(commits):
 next_index = index + 1
 if next_index < len(commits):
 commit_diffs[commit.hexsha] = repo.git.diff(commits[next_index], commit)
 return commit_diffs