dupe-krill VS juniper

Coincidentally I was looking through the list of apps that do this the other day, I think I looked at all the other ones they list as competitors.

My particular use case involved images that might be very near duplicates (screenshots of the same web page), which some of the tools cover, though it feels like a slightly seperate task from the exact bit duplicates, so not all do it.

One interesting one I found that wasn't listed in the Readme was:

https://github.com/kornelski/dupe-krill

Which had some notes about their use of BTreehashes to progressively compare files. Not sure how much difference it makes in practice but sounded elegant.

juniper

GraphQL is also an option. https://github.com/graphql-rust/juniper

Juniper has plenty of popularity, so I guess it's Rust + graph DB that doesn't get much love then... :/

rusty-gql would not be able to release without async-graphql and juniper.

I'm looking for graphql server that can do queries and mutations over websocket, like subscriptions-transport-ws. Juniper and async-graphql both looks promising and async-graphql at least uses wording Subscriptions (WebSocket transport) in features but i couldn't find much more or any examples about that from the docs or repo.

We also don’t use Node anymore. We found Node in general to be incredibly slow, not to mention single threaded. Admittedly, we did use Node (Apollo) though up until about two years ago. Now, depending on use-case, we will use Go (https://github.com/99designs/gqlgen) or Rust (https://github.com/graphql-rust/juniper). Both outperform a Node gateway but significant margins.

(example from the docs async-graphql https://async-graphql.github.io/async-graphql/en/subscription.html, juniper: https://github.com/graphql-rust/juniper/blob/master/docs/book/content/advanced/subscriptions.md)