Getting the full abspath instead of the basename from a Blob

[daveaitel]

So from what I can tell there is no way to get the full abspath (aka, like what you would get from git diff-tree --name-only -r for modified files.

This is not great in the sense that when gathering metadata on a project, you often have name collisions (aka, Makefile can exist in two directories). From what I can tell tell, the ModifiedFile object gets initialized from the GitPython object's new path, and getting abspath would not be trivial. :(

Thanks,
Dave

[ishepard]

Hey Dave!
Mmm the names should be different, right?
I mean, if I have:

- repo/
 - file1.txt
 - dir1/
     - file1.txt

and I change both files, I will have:

file1.txt
dir1/file1.txt

so you shouldn't have conflicts, right?

[daveaitel]

This would be the ideal, but actually what you get is just: file1.txt and file1.txt :) In domain/commit.py: self._new_path = Path(new_path) if new_path is not None else None From parse_diff: def _parse_diff(self, diff_index) -> List[ModifiedFile]: modified_files_list = [] for diff in diff_index: old_path = diff.a_path new_path = diff.b_path Which I believe comes from: https://github.com/gitpython-developers/GitPython/blob/main/git/diff.py Which if I understand correctly, is the final blob's name (aka, not a Path.join(tree[0:]) as you would want for abspath?

…

-dave

On Wed, Jun 23, 2021 at 9:33 AM Davide Spadini ***@***.***> wrote: Hey Dave! Mmm the names should be different, right? I mean, if I have: - repo/ - file1.txt - dir1/ - file1.txt and I change both files, I will have: file1.txt dir1/file1.txt so you shouldn't have conflicts, right? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#188 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE25MYSAK7GPOOPTZ2IW66TTUHPEHANCNFSM47FULRNQ> .

[ishepard]

Mmmmm no then there is something wrong :) you get the full path.
Try to run:

from pydriller import Repository

for commit in Repository("test-repos/complex_repo").traverse_commits():
    for mod in commit.modified_files:
        print(mod.new_path)

and you will see:

Matricula.java
Aluno.java
Matricula.java
Secao.java
Secao.java
None
Matricula.java
Matricula.java
Matricula.java
Matricula.javax
Secao.javax
pasta/Capitulo.java.    <-------- as you can see, there is the full path
.DS_Store
Arquivo.java
Arquivo.java

[daveaitel]

Is modification.filename different from modification.new_path ? Perhaps that is what I am failing at? * small snippet of my code here * for modification in commitfiles: print('Author {} modified {} in commit {}'.format( commit.author.name, modification.filename, commit.hash)) *Thanks for the responses btw.

…

On Wed, Jun 23, 2021 at 11:20 AM Davide Spadini ***@***.***> wrote: Mmmmm no then there is something wrong :) you get the full path. Try to run: from pydriller import Repository for commit in Repository("test-repos/complex_repo").traverse_commits(): for mod in commit.modified_files: print(mod.new_path) and you will see: Matricula.java Aluno.java Matricula.java Secao.java Secao.java None Matricula.java Matricula.java Matricula.java Matricula.javax Secao.javax pasta/Capitulo.java. <-------- as you can see, there is the full path .DS_Store Arquivo.java Arquivo.java — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#188 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE25MYRJPFQ7REKVQK57LHDTUH3UFANCNFSM47FULRNQ> .

[ishepard]

Oh mistery solved :)
See here: https://pydriller.readthedocs.io/en/latest/modifiedfile.html

• filename: return only the filename (e.g., given a path-like-string such as “/Users/dspadini/pydriller/myfile.py” returns “myfile.py”)
• new_path: new path of the file (can be None if the file is deleted)
• old_path: old path of the file (can be None if the file is added)

What you want is the path (like in git, you have 2 paths, old and new, before and after the change, so you need to check both).

No worries, glad I could help!