Skip to main content

JSON serialization support for generated Scala classes

The purpose of this documentation is to give an explanation on how the JSON Codec Backend works and how to use it. This backend is a feature built on top of the Scala backend, which allows you to read Scala types that the Scala backend generates from JSON and write to JSON.

How to use the JSON Codec Backend​

The first step to using the JSON Codec backend is to add the Circe JSON library and morphir sdk as dependencies to your project. The morphir sdk can be added to your project through the following ways:

  1. Copy dependencies from morphir-jvm : This option allows you to copy the dependencies(morphir sdk) used by the generated code to the output path. You can achieve this by adding the -c or --copy-deps flag to the morphir-gen command.

  2. Adding from Maven Repository : The SDK can be added from the maven repository as a dependency . It can be found at Morphir SDK Core.

Generating Codecs from the IR​

The morphir-ir.json is a prerequisite to generating codecs. Make sure you have that and proceed with the following steps

  1. Run morphir-elm gen --include-codecs --target Scala to generate the codecs
  2. In the situation where you do not have morphir sdk as a dependency in your project, you have to add the --copy-deps flag to copy dependencies from the morphir-jvm project. Complete command for this is morphir-elm gen --iclude-codecs --copy-deps --target Scala.

How to use Generated Codecs​

Using an example , lets describe a model structure and look at how the generated codec for this particular model will be structured

The model:

- reference-model
  - src
    - Morphir
      - Reference
        - Model
          - BooksAndRecords.elm  
- morphir.json
- morphir-ir.json

Running the command morphir-elm gen --include-codecs --copy-deps --output codecs/src to generate codecs means your codecs would be outputted into the codecs/src folder in the structure:

- reference-model
  - codecs
    - src
      - morphir
        - reference
            - model
              - booksandrecords
                - Codec.scala   <----- This is your genereated codec 
- morphir.json
- morphir-ir.json

The generated codecs Codec.scala has both encoders and decoders.

Importing and Using generated Codecs​

In the example above the codecs where generated based on the BooksAndRecords.elm model and output in the codecs package. As an example lets say we want to use the generated codecs in a package called usecodecs with the following structure:

- usecodecs
  - src
     - UseCodecs.scala

Inside UseCodecs.scala we need to import the codecs from the package in which it was generated ie: morphir.reference.model.booksandrecords and use the codecs from the Codec object. Below is an example of how to export and use a codec.

import morphir.reference.model.booksandrecords.Codec._
import io.circe.Json._
import io.circe.parser._

object UseCodecs {
    def main(args: Array[Strings]): Unit = {
        val bookType = BookType("HardCover")
        val encodedBookType = encodeBookType(bookType)      <--- Using encoder
        
        val parseBookType = parse(encodedBookType.toString()).getOrElse(Null)
        val decodedBookType = decodeBookType(parseBookType.hcursor)        <--- Using decoder
    }
}

It is required to import io.circe.Json in the situation where you want to use methods from the circe library. Libraries such as io.circe.parser and io.circe.generic are usually used to work with encoders and decoders, you would need to add any additional library as a dependency to your project in order use them with the codecs.

Unsupported Features​

Codecs are not generated for function types in the morphir ir.