We’ve worked on it at the last hackathon and few days after at home with @poka.
List on what’s have done :
Docker
- build docker image of the indexer itself and another one for Hasura with migration
- configuring the CI to auto publish docker images
- refactoring docker-compose with
prod
,dev
andtest
mode- prod : launch all pre-configured services
- dev : launch all needed services, but not the indexer
- test : launch duniter-rpc with sealing mode
Hasura
- bump Hasura version
- refactoring database :
- with name as the primary key of identity table
- renaming block index
- renaming id to pubkey
Fastify
- fill database directly with real sql queries and not through graphQL : more than x2 faster!
- refactoring the process block mechanism with an event emitter
- remove
fastify-hasura
andvite-node
dependencies, which simplify docker workflow - started writing test with Vitest and fake data with Polkadot
- indexing the true Ğ1 genesis and all transactions !
Frontend
- designed awesome logo
- refactored frontend and integrate anu-vue UI lib
- update graphiQL to the latest
- toggle mode relay graphQL API
- display latest finalized block
You can take a look online here :
https://gdev-indexer-graphiql.p2p.legal