bee-produced / bee-built Goto Github PK

Nested data fetchers that only call data loaders could be generated automatically to reduce boilerplate code. If they are generated automatically, one can also not forget to define them.

Example for such data fetchers:

@DgsData(parentType = "Film", field = "studios")
fun filmStudios(dfe: DataFetchingEnvironment): CompletableFuture<List<CompanyDto>?> {
    val dataLoader = dfe.getCompanyDataLoader()
    val film = dfe.getSource<FilmDto>()
    return dataLoader.loadMany(film.studioIds)
}

@DgsData(parentType = "Film", field = "directors")
fun filmDirectors(dfe: DataFetchingEnvironment): CompletableFuture<List<PersonDto>?> {
    val dataLoader = dfe.getPersonDataLoader()
    val film = dfe.getSource<FilmDto>()
    return dataLoader.loadMany(film.directorIds)
}

@DgsData(parentType = "Film", field = "cast")
fun filmCast(dfe: DataFetchingEnvironment): CompletableFuture<List<PersonDto>?> {
    val dataLoader = dfe.getPersonDataLoader()
    val film = dfe.getSource<FilmDto>()
    return dataLoader.loadMany(film.castIds)
}

One could in theory do this at runtime, however this would require an approach heavy reflection usage. In addition, the DGS plugin would not indicate in the schema that a data fetcher is defined.

Streamline project structure of older modules:

• Consistent naming scheme (bee)
• Move functional capabilities to module itself (@TransactionalResult should become part of lib.data)
• Extensions functions should be placed in packages where the name is of the type which is extended
• AutoConfiguration for Spring Libraries (e.g. EventManager)
• Bump versions

Also provide some minor improvements for lib.data:

• Add support for typedSelection which was internally in development but is now nearly finished
• Add extension dfe.toDataSelection()

Test & prototype implementation with Blaze Persistence instead of Entity Graph

The developers of Blaze Persistence see entity views as a better alternative to JPA 2.1 entity graphs...

• Test Inheritance

• Test Updates

• Test Transactions

  • Query again after update?
  • Insert data after transaction begin?

• Test Filters

  • Is it possible to add kotlin jdsl?
  • evm.applySetting(setting, cbf.create(...).also { // add jdsl })

• Codegen Views

  • As Blaze does not allow recursive views generate copies until threshold e.g. 2
  • Analyze schema or DTOs?
  • How to model skip over?
    • Model in schema?
    • How to omit m:n tables in schema?
    • Custom GraphQLEntityViewSupport

• Codegen Selection

  • Selection should be possible without GraphQL DataFetchingEnvironment!

    • Custom GraphQLEntityViewSupport

  • Remove FullNonRecursiveSelection => Generate Full Simple Selection for every Type

  • Generate Simple Selection DSL

    select(Show.DSL {
    	actors {
    		firstname()
    		lastname()
    		companies {
    			users()
    		}
    	}
    })

• Generate real data classes instead of bytebuddy proxy

  • See TupleConstructorReflectionInstantiator

Add existing in-house libraries to project & implement example project showcasing features.

Create pipeline with GitHub actions that tests libraries.

Also publish artifacts to Maven Central.

Implement better solution to improve compatibility between result type & data fetchers in bee.functional.dgs.
Investigate ByteBuddy for possible solution.