Getting Started Project - automatic cache followed, tags, &c

Hi,

I’ve got an idea for a project that might be useful, that I think I can do with a couple of starter tips.

  1. I’ve got what appears to be a functional lbrynet on a headless amazon server (no X, just SSH).
    I can claim search --channel @LibertyDoll.
    Then, parse json, pull out the permanent url, and get it.
    At that point, that’s all I have to do to both have a copy and become a source for others to pull from, right?

  2. So, basically, I’d like to write something that runs as a service (no problem), listens or periodically queries for new claims by poster, or tag or tag/combination, and today’s new, trending, popular, and manages the disk space of each (i.e. keep everything from @LibertyDoll, anything from Akkad that has more than 5 lbc in support, and all my gun channel stuff for a month times the number of LBC it’s supported with. As you get closer to disk limtations, priority remove various stuff, keeping around metadata.) Maybe in the future, do something like if I deign to comment on a video, keep it around for posterity.

  3. My preferred language is rust, but if there’s no libraries in it that language, I can do python, or just script lbrynet command output via json parse, whatever.
    Any suggestions in these directions would be welcome.

  4. Any other suggestions would be valuable.

I want to make sure that, at least, I can get at things of interest, and provide duplicate storage for those people I follow/respect. Also, my inclination would be to have some way of biasing my desktop towards trying to get the blobs from my personal instances, have bandwidth constraints for sharing that are generally in force but that I can violate because I’m cool (so, outside of any plebian bandwidth constraints, I can still get full bandwidth/class of service), but that might not be an easy thing.

(Maybe caching the metadata for new, trending, and top, and we can run ML on it, so the AIs, too, can learn hatred and contempt of bitcoin scammers.)

  1. Finally, where’s a good repo to put this up? Should I just make a random github, and then link to it from some dev page? Are there useful naming conventions that might apply? lbrycache seems like it should be a thing.

Looking forward to working on it, should be fun!