| .gitignore | ||
| .python-version | ||
| default.nix | ||
| git_annex_duplicates.py | ||
| LICENSE.txt | ||
| pyproject.toml | ||
| README.md | ||
| uv.lock | ||
git-annex-duplicates
Find duplicate files in a git-annex repository.
git-annex-duplicates groups files by their annex key and prints every group
that contains more than one path. It is read-only: it reports duplicates but
does not remove or otherwise modify any files.
Requirements
- git-annex
- Python 3.11 or newer
- uv (recommended for installation)
Installation
Install the command directly from the Git repository:
uv tool install "git+https://cvs.moegen-wir.net/mikegerber/git-annex-duplicates.git"
Alternatively, run it from a checkout:
git clone https://cvs.moegen-wir.net/mikegerber/git-annex-duplicates.git
cd /path/to/your-git-annex-repo
uv run --project /path/to/git-annex-duplicates git-annex-duplicates
Usage
Run the command from inside a git-annex repository:
git-annex-duplicates
This will list any duplicate files below the current working directory. The output contains one group of duplicate paths at a time, with groups separated by a blank line:
photos/holiday.jpg
backup/holiday.jpg
documents/report.pdf
archive/report-final.pdf
Exclude one or more path patterns with --exclude:
git-annex-duplicates --exclude '*.iso' --exclude 'archive/**'
Find duplicates in the given directories:
git-annex-duplicates dir_a/ dir_b/
Finding duplicates across directories
To find duplicates that are duplicated across directories, use the --across flag:
git-annex-duplicates --across data/ new_data/
This will only list duplicate files that are found in more than one of the given directories.
How it works
The command runs git annex find, requests each file's annex key, groups the
returned paths by that key, and prints groups containing at least two paths.
Files with the same annex key refer to the same content as far as git-annex is
concerned.
Only annexed files are considered. Regular Git files and untracked files are not included, and file content does not need to be downloaded to produce the report.
Contributing
This program is almost trivial, so I don't expect many contributions. However, if you wish to implement a new feature, please open an issue at GitHub first, so we can discuss it.