meilisearch-demo

Trying out MeiliSearch, a lightweight open-source search engine, paired with a client written in Java (Spring Boot).

Setup

• run docker-compose up meilisearch to spin up Meilisearch instance
• master key for Meilisearch is masterKey for development purposes
• calling Meilisearch instance through REST calls use done using Authorization: Bearer masterKey header
• you possibly need to change spring.sql.init.mode property to seed the schema once

What this does

• Allows user to create an entity through POST /api/project endpoint
• The creation of this entity gets intercepted and results into a job being enqueued
• The job gets the persisted entity (projected as a DTO) and persists this into Meilisearch under the index name of MyFancyDto.class.getSimpleName()
• The GET /api/project endpoint returns the stored dtos in Meilisearch

General idea

• Have some kind of background syncing to other datastores (which are faster for certain purposes) while not disturbing the write performance.
• Concrete implementations of this idea could be persisting an audit trail or using the persisted data for a dashboard that does not need 'realtime' info

TODO / future improvements

• Extract specific entity fetching out of job. Would rather like to see a handler per entity, which get activated in the job.
• Add specific searching (filtering? sorting?) capabilities to GET endpoint
• Store metadata about the Meilisearch sync onto the entity or on the DTO in Meilisearch