so i mentioned working on a library for working with activitypub servers a few days ago and writing it has been going well so i figured i’d share what it is and where it’s at.
the library is called (tentatively) “activitysearch” and it operates around the model of a single actor. an actor has an associated database (pouchdb) where it stores any data it downloads about the actor — but it need’s that actor’s permission to get the necessary creds. it is not a tool for checking out other people on the fediverse. it is a tool for working with your own data.
the data it downloads is then indexed so you can make queries against it without hitting the network, queries like, “who am i following whose instance has disappeared?” or “where’s that toot i fav’d about crows?” i have a rudimentary text-search algorithm i use to build a search index but it’s not really professional-grade like lucene, just something useful enough for general purposes.
re: activitysearch preview
right now i’m wrapping up the testing phase, specifically doing integrity testing to ensure the library handles its data as intended (ex: doesn’t mysteriously delete or duplicate things) but otherwise everything works and you can use it for what it says on the tin. after testing, i’ll write the docs and publish the source.
i’m planning a CLI, browser app, and self-hosted server for this model of handling your own data. after that, well, i’m thinking about writing apps that treat AP as just one of many media channels, including things like secure-scuttlebutt and dat’s budding social media apps. but who knows what the future holds 🤷♀️