Jacinta

Simple interfaces for reading, processing and writing JSON

Jacinta is a fully-featured JSON library for Scala, built upon the JSON parser, Merino, and designed to make it easy and safe to work with JSON in Scala.

Features

• parse and serialize JSON
• intuitive dynamic API for quick field access, without compromising typesafety
• automatic conversion to and from product and sum types

Availability

Getting Started

All Jacinta terms and types are in the jacinta package, and exported to the soundness package. So we begin either by importing,

import jacinta.*

or:

import soundness.*

Core types

Jacinta's most important type is Json which represents an instance of a JSON value, that is, a JSON object, array or primitive (string, number or boolean). It does not represent serialized JSON, so details like whitespace and the ordering of keys in an object are not represented.

Scala's type system knows nothing more about the internal structure of a Json value than this. So a Json value representing the number 12 is indistinguishable by the type system from a Json value representing an array of complex objects.

Parsing

We can obtain a Json value by constructing one from existing values.. Or we can parse some textual input.

The Json.parse method takes any input that is Readable by Bytes. This includes not only Text and Bytes values, but other types like filesystem Paths—if suitable context is provided for a Readable by Bytes value to be resolved.

Here is an example of parsing Text as JSON:

Json.parse(t"""{ "name": "Alfred", "age": 83 }""")

Calling Json.parse may raise a JsonParseError, so this should be handled in some way. Full details of error handling in Soundness is provided by Contingency.

Dynamic field access

As a dynamically-typed value, Scala's type system does not know anything about the fields that are available on a particular Json value. It does not even know if it is an object with fields, an array with indices, or a primitive value.

But we may know more than the type system. Or at least, we may wish to program to the assumption that we know more. So Jacinta makes it possible to access fields and array indices dynamically using selection or arbitrary field names with the familiar ., and application with parentheses for numeric indices.

This would normally be a significant compromise on typesafety, since it would allow us to call nonexistent methods on Json values, without protection from the compiler. So access must be explicitly enabled with the import:

import dynamicJsonAccess.enabled

With this contextual value in-scope, we can dereference and deindex Json values, dynamically. The result will always be another Json value, ready to be deindexed or dereferenced, or a JsonError will be thrown if the index is out of range, or the object key does not exist.

Json values are not useful (in most cases) for use elsewhere in Scala code, unless we can convert them to typed Scala values. This is called decoding.

Decoding Json values

Encoding as Json

Many Scala values can be mapped directly (and often unambiguously) to JSON values. Trivially, this includes Text, Boolean and various numeric types. But collection types like List and Set can also be mapped to JSON arrays, if their elements are types which can be mapped. And case classes and tuples of these types may also be mapped, so long as their elements can. With a suitable choice of encoding, sum types (like enumerations or sealed traits) can also be mapped.

This is true compositionally. For example, Lists of sealed traits, composed of case classes whose parameters are tuples of Ints, Sets and other case classes are equally encodable.

Encoding a value to Json is as simple as calling .json on that value. If it is able to be encoded, the result will be a Json value. (Note that this is not the same as encoding a Json value to string-like representation, which is a useful—but different—operation, described below.)

Here is an example of a simple, but nontrivial case-class structure:

case class Person(firstName: Text, lastName: Text)
case class Recipient(person: Person, emailAddress: Text)

val recipient = Recipient(Person(t"Piotr", t"Nowak"), t"pn@example.com")

Given these definitions, the recipient instance can be encoded with recipient.json into a Json value representing the following JSON object:

{
  "person": {
    "firstName": "Piotr",
    "lastName": "Nowak"
  },
  "emailAddress": "pn@example.com"
}

Coproducts

Encoding as Text

Decoding as Json

Sealed traits of two or more case class subtypes will be serialized to JSON objects, exactly as each of the subtypes would be, but with an additional field called _type, whose value will be set to the unqualified type name, e.g. "_type": "Leaf".

Although this encoding of coproduct types is non-standard, it is a reasonable default, and can always be overridden with specific typeclass instances for the sealed trait type.

Collections

Furthermore, all traversable standard collection types can be serialized to JSON arrays, provided the elements of the collection can be.

Acessing values

Instances of Json are dynamically-typed which means that members with arbitrary names may be accessed as if they were methods. Taking the recipient example above, it would be valid to access recipient.person, as if the method person existed on the Json type. It doesn't, but since Json instances inherit from the special Dynamic trait, the code will be transformed into recipient.selectDynamic("person") at compiletime, which will return a new Json instance representing the JSON:

{
  "firstName": "Mike",
  "lastName": "Smith"
}

It is therefore possible to call recipient.person.firstName directly and get a Json value representing the JSON string, "Mike".

As dynamic values with little known statically about them, instances of type Json are not particularly useful directly, and should be converted to other types like Text, Int or Person before being used elsewhere in a program. This is achieved with the Json#as method which takes the destination type as a parameter, for example,

val addressee: Text = recipient.person.firstName.as[Text]

or,

val person: Person = recipient.person.as[Person]

As well as accessing arbitrary fields in a JSON object, elements of an array may be accessed by simply applying the integer index to a Json value representing an array, for example, json.organisation.users(2).as[User].

Errors

Since dynamic field access is unchecked at compiletime, it's possible that a JSON object would not contain the requested field, or a JSON array would not contain the requsted index. This will throw an exception only when attempting to convert the value to a static type. So the expression, recipient.user.firstName, (noting that user is not a valid field of recipient) would not produce an error in itself. Only when invoking, recipient.user.firstname.as[Text] would an exception be thrown, of type JsonAccessError.

Similarly, if the expression, recipient.person.firstName.as[Int] were evaluated, a JsonAccessError would be thrown due to the field, firstName being a JSON string and not a JSON number.

All methods which throw exceptions are annotated with throws clauses, and if saferExceptions is enabled, these must be handled.

Typeclasses

While all Java primitive types and Strings, collection types and case class types can be serialized and deserialized automatically, it's possible to support other types or to replace existing default implementations by providing contextual instances of the typeclasses, Json.Writer and Json.Reader.

For example, assuming the existence of an Email type (which simply wraps a Text instance), a Reader and Writer for Email could be provided in Email's companion object, like so:

case class Email(value: Text)

object Email:
  given Json.Reader[Email] = json => Email(json.as[Text])
  given Json.Writer[Email] = _.value.json

Note that Email is a case class, so default instances of Json.Reader[Email] and Json.Writer[Email] would exist already, but would be replaced by these new definitions. (If Email were instead a non-case class, these would be chosen unambiguously as the only contextual instances.)

Functor and Cofunctor

Json.Readers are functors, and the Reader#map method is provided to transform a reader of one type into a reader of another. Likewise, Json.Writers are cofunctors with Writer#contramap methods. Given these definitions, an alternative way to write the definitions for Email by transforming the existing instances for the Text type would be:

object Email:
  given Json.Reader[Email] = summon[Json.Reader[Text]].map(Email(_))
  given Json.Writer[Email] = summon[Json.Writer[Text]].contramap(_.value)

Status

Jacinta is classified as fledgling. For reference, Soundness projects are categorized into one of the following five stability levels:

• embryonic: for experimental or demonstrative purposes only, without any guarantees of longevity
• fledgling: of proven utility, seeking contributions, but liable to significant redesigns
• maturescent: major design decisions broady settled, seeking probatory adoption and refinement
• dependable: production-ready, subject to controlled ongoing maintenance and enhancement; tagged as version 1.0.0 or later
• adamantine: proven, reliable and production-ready, with no further breaking changes ever anticipated

Projects at any stability level, even embryonic projects, can still be used, as long as caution is taken to avoid a mismatch between the project's stability level and the required stability and maintainability of your own project.

Jacinta is designed to be small. Its entire source code currently consists of 618 lines of code.

Building

Jacinta will ultimately be built by Fury, when it is published. In the meantime, two possibilities are offered, however they are acknowledged to be fragile, inadequately tested, and unsuitable for anything more than experimentation. They are provided only for the necessity of providing some answer to the question, "how can I try Jacinta?".

1. Copy the sources into your own project

   Read the fury file in the repository root to understand Jacinta's build structure, dependencies and source location; the file format should be short and quite intuitive. Copy the sources into a source directory in your own project, then repeat (recursively) for each of the dependencies.

   The sources are compiled against the latest nightly release of Scala 3. There should be no problem to compile the project together with all of its dependencies in a single compilation.

2. Build with Wrath

   Wrath is a bootstrapping script for building Jacinta and other projects in the absence of a fully-featured build tool. It is designed to read the fury file in the project directory, and produce a collection of JAR files which can be added to a classpath, by compiling the project and all of its dependencies, including the Scala compiler itself.

   Download the latest version of wrath, make it executable, and add it to your path, for example by copying it to /usr/local/bin/.

   Clone this repository inside an empty directory, so that the build can safely make clones of repositories it depends on as peers of jacinta. Run wrath -F in the repository root. This will download and compile the latest version of Scala, as well as all of Jacinta's dependencies.

   If the build was successful, the compiled JAR files can be found in the .wrath/dist directory.

Contributing

Contributors to Jacinta are welcome and encouraged. New contributors may like to look for issues marked beginner.

We suggest that all contributors read the Contributing Guide to make the process of contributing to Jacinta easier.

Please do not contact project maintainers privately with questions unless there is a good reason to keep them private. While it can be tempting to repsond to such questions, private answers cannot be shared with a wider audience, and it can result in duplication of effort.

Author

Jacinta was designed and developed by Jon Pretty, and commercial support and training on all aspects of Scala 3 is available from Propensive OÜ.

Name

Jacinta is one of the feminine forms of the given name Jason, which is homophonous to JSON.

In general, Soundness project names are always chosen with some rationale, however it is usually frivolous. Each name is chosen for more for its uniqueness and intrigue than its concision or catchiness, and there is no bias towards names with positive or "nice" meanings—since many of the libraries perform some quite unpleasant tasks.

Names should be English words, though many are obscure or archaic, and it should be noted how willingly English adopts foreign words. Names are generally of Greek or Latin origin, and have often arrived in English via a romance language.

Logo

The logo shows a pair of braces—important syntax in JSON—positioned to look like two people face-to-face, alluding to the concept of communication.

License

Jacinta is copyright © 2025 Jon Pretty & Propensive OÜ, and is made available under the Apache 2.0 License.