🗓️ The preliminary #distribits2025 schedule is now available: https://www.distribits.live/news/2025-distribits-schedule/.

You can expect to hear talks about federated data management across disciplines, deep dives into :gitannex: #gitAnnex and special remotes, how to make your analysis pipelines #FAIR from beginning to end (and tools you can use to make it happen), and so much more.

If you missed the registration deadline but would still like to participate, stay tuned - we will be able to accommodate late registrations soon.

There is now also a #gitAnnex wheel on PyPi that should run on any system with GLIBC 2.17+. This should work on CentOS7 (released more than 10 years ago).

If you have access to such a machine, please consider giving it a try.

Thanks!

Is there a tool to set up #GitAnnex to calculate archive unpackings yet? Something™ could look at a repo's .zip / .tar.gz files, check whether any file in the repo is also in there, and annotate it as computed / computable. (I'm not aiming to restore the zip files from the content –that'd be pristine-tar territory. I'd already be happy if the compute path would be there so that the single files don't need to be in the archive).

Yay, @joeyh made it so :gitannex: #gitAnnex pushes the current branch first (not the git-annex or synced/* branches). This is awesome when you want to push-to-create a local repo to your @forgejo or #GitLab instance. Previously, `git annex assist` et al. would result in e.g. synced/main (or worse: synced/git-annex) being used as the default branch on the forge, which is never what you want.

https://git-annex.branchable.com/todo/Pushing_current_branch_before_synced__42___or_git-annex/?updated#comment-3eb491fcc789d056141c9d388d15819f

As of version https://pypi.org/project/git-annex/10.20250520b7 the #gitAnnex wheel is now tested and confirmed to be functional after a

`uv tool install git-annex`

on Linux (tested with Ubuntu 22.04 and 24.04), Windows (tested with 2022 and 2025), and Mac (tested with 13(intel), 14(arm), 15(arm)).

https://github.com/psychoinformatics-de/git-annex-wheel/actions/workflows/test-pypi-wheel.yaml

There is now a #gitAnnex package on #PyPi: https://pypi.org/project/git-annex/

This should make it simpler to deploy git-annex in Python virtual environments, also as versioned dependencies for software like #Datalad

Packages are built for Linux, Windows, and Mac via GitHub actions: https://github.com/psychoinformatics-de/git-annex-wheel/

Contributions to cover more platforms are most welcome!

@alerque There's a discussion here¹ between me and @Atemu. They prefer #paperless' tagging and just want to have the files in :gitannex: #gitAnnex as well, while I'd rather use paperless as an extension for quick search etc.

¹https://fosstodon.org/@nobodyinperson/113789283043978309

Has anybody tried combining #gitannex with #paperless ngx? I use the former extensively for other purposes, but want some additional tooling to layer with it.

I have a genre of paper documents that I have had locked up in a proprietary app. I thought my use case was going to be short lived and at the time the tooling for managing scans was far superior in the free tier of a commercial solution, but I've definitely outgrown it and want to get this genre of my data under my own control again.

I don't know why, but on #NixOS 25.05 I'm back again keeping my worktree nice and DIRTY so I can use :gitannex: #gitAnnex'ed files. 🙄

@joeyh considers removing the :gitannex: #gitAnnex webapp as probably only few people use it, it hasn't seen significant changes in a long time, has potential security implications and increases git-annex's size.

More info here: https://git-annex.branchable.com/todo/remove_webapp/

Do you use the git annex webapp?

:boost_requested: #boostsWelcome

@mmeier @timo Forgejo also has a fork with #gitAnnex support to store large files better than git LFS. I guess some day it'll land in forgejo proper.

How to properly re-setup #gitannex repositories?

I replaced my paperless-ngx workflow with #denote plus #gitannex for archiving arbitrary files (as denotes naming scheme can applied to those as well) like invoices, pdf documents etc. and tagging them. I use rg.el with rg-executable set to ripgrep-all for full text search so that's another thing solved. This works exceptionally well for me so far & means there's one less web application to host.

Having a ball doing #literatureResearch with @zotero. I previously hated doing it, but it turns out it was because I didn't have the right tool.

I used #kbibtex and :gitannex: #gitAnnex before. Robust, decentralized, very fine syncing control, fully offline-capable, etc. - but it was missing the streamlined, well-integrated polish that #Zotero has and felt a bit too clunky to be efficient.

#PhDLife #academicMastodon

Oh wow and C++ #nix (now?) apparently doesn't mind :gitannex: #gitAnnex anymore. #lix for example refuses to do anything if the flake.lock is behind a smudge filter and your worktree is clean. You have to *manually* dirtyfy the worktree to be able to work with lix. Multiple good arguments to stay on #cppnix for me.

My daughter prettified my #EGU25 poster, I love it 😂

Come find me in hall X5 poster X5.166 talk about our copter measurements and how I use :gitannex: #gitAnnex to manage the data.

#UAS #multicopter #meteorology #PapaSein #parenting #FediEltern

@nobodyinperson ah, from what I understand, this is metadata that is stored in the git-annex branch, so it should be replicated... 🤔

I never worked with `wanted` and metadata in #gitAnnex at all.. have to ensure I get things right.

Asking the #gitAnnex master here on mastodon: @nobodyinperson - can you configure a git repository to be a git-annex repository but never accept data?

Usecase: I have a file on my homeserver which is a encrypted disk that contains git repositories (config, password-store, etc) and I want to have the git-annex repository on there, but never the actual data because that filedisk is only very small - I just want to have metadata on there, never the actual blobs.

Is this possible and if yes, howto?

My first guess would be using a bare git repo and just doing `git push --all` and never `git annex sync` - but maybe theres a better way?

@matrss and @mih this might interest you 👆

I wonder how it plays with :gitannex: #gitAnnex, but I do have several different diffdrivers for annexed files (plaintext, jupyter notebooks, etc.) so it should work I guess?

@khinsen Great to see #gitannex on the curriculum and our @datalad listed among "fancy projects" in the promotional video! Shameless plug: you might be interested to discover/participate-in/present at @distribits .