Wasn't satisfied with zotero and mendeley for pdf reference management. Wanted something terminal based and simple, so I wrote a shell script that I eventually rewrote in python. Given a pdf, it reads it, tries to find the article doi and fetches metadata from crossref to rename the file and store metadata in the file itself. I Plan on integrating features like interactive tagging and tag browsing from the terminal.
Any help with code review or additional features are welcome :)
This is ultimately due to the general problem of neovim internals being in so much flux. So much keeps changing, updating, and breaking in the internals and plugins. It does take some upkeep at every level, and after a point I think people get tired of it. At the end of the day, you only have so much time and patience.
I had written a script to manage my pdf literature using crossref and arxiv to fetch metadata and store them in pdf tags, and just renaming the pdf in place. No hidden complexity, no gui.
I hacked it into a python package a couple of months ago in an ADHD hyperfocus induced frenzy. Sounds like you might like it. Still a long way to go, but it's out there and FOSS. Contributions are welcome.
Nix has potential to be a universal installer, although it's still a bit cumbersome when working with languages like python. Another pain point is the need for superuser privilege to write to /nix
My biggest gripe with nix is that you need sudo rights to get it running properly on another distro. I know you can do the nix-user-chroot thing, but it's not possible on machines where kernel namespaces aren't accessible for unprivileged users.
I spent time understanding it, and getting it working, but it really doesn't work just universally. Once the install process becomes more like, say, miniconda -- where an unprivileged user can install it locally with ease -- nix will get a massive boost in my books. Until then, I can only yearn.
It is possible to use from a completely unprivileged environment by putting the nix store in your home directory instead of `/nix`. But (and it's a big but) you'll then have to build all packages from scratch because the hashes for you packages will be different to the ones in the nix binary cache.
This is the second comment in the entire thread. And the poster is just one person. What makes you generalize something to an entire community, seemingly without any basis?
It's about creating a habit, like many people mentioned here. For me, the git commandline process was a bit too much hassle and typing to do if I wanted to do it atomically. In the end, I wound up using lazygit[0] and the nvim plugin[1] to make life easier. It makes committing only selected chunks a lot easier, not to mention easier cherry picking and other stuff.
