That's great!
What is the tl;dr on NIP-85?
I look forward to seeing your product including your DVM-NIP. I could easily offer PageRank, GrapeRank, follower lists, and all sorts of other scores by API to devs because it’s an easy thing to do, once the graph db is running.
But API was never my end goal. I love nostr:npub1gcxzte5zlkncx26j68ez60fzkvtkm9e0vrwdcvsjakxf9mu9qewqlfnj5z ‘s suggestion of publishing the data using NIP 85, kind 30382 events.
Discussion
I’m going to publish one kind 30382 event per pubkey (180k or so) which includes their various WoT scores and metrics, as determined from my pubkey’s perspective (bc all reputation is relative). The proposed NIP-85 includes tags for Follower count as well as a generic “rank” from 0 to 100 which I suppose could mean anything — I’m thinking I’ll adapt the GrapeRank influence score, which ranges from 0 to 1.
I may add other tags … maybe one for PageRank score? … although that’s not currently in the NIP.
If it's all from your perspective, what is the usecase here? Or you want to store all possible permutations? Sounds like a lot of events
One step at a time. I’m currently working on a server that runs strfry + neo4j (already got that working), which imports data from LMDB to neo4j (currently working on the ETL pipeline for that), calculates PageRank, GrapeRank and other scores, and uses them to manage the filters that strfry uses when it imports from other relays. This will be the next-gen personal WoT relay.
So we take things one step at a time.
Step 1: you will go to Amethyst, view Alice’s profile and see kind 30382 event data authored by my server, that is already functioning today. Which scores will users want? PageRank? GrapeRank? Follower count? *Validated* follower count? Something else? We will need users to tell us!
Step 2: same thing, except that the kind 30382 events are authored by YOU and stored on YOUR personal WoT server.
Thinking I’ll pair neo4j with nostr:npub1fvmadl0mch39c3hlr9jaewh7uwyul2mlf2hsmkafhgcs3dra6dzqg6szfu as the basis for next-gen Personal WoT Relay.
First step will be some scripts that synchronize the relay’s LMDB with neo4j. Looks like the node-lmdb library may do the trick.
Second step will be scripts that manage the graph db as your own personal Nostr Knowledge Graph.
Third step will be calculation of PageRank, GrapeRank, and other WoT-related metrics.
These tools will be readily adaptable to any relay that uses LMDB, which includes strfry, khatru, and also nostrdb (notedeck) if I am not mistaken.