This is a small library for parsing YAML into circe's Json AST.

• For parsing YAML 1.1 it uses SnakeYAML.
• For parsing YAML 1.2 it uses snakeyaml-engine.

YAML is a useful data format for many purposes in which a more readable, less verbose document is desired. One use case, for example, is human-readable configuration files.

SnakeYAML provides a Java API for parsing YAML and marshalling its structures into JVM classes. However, you might find circe's way of marshalling into a Scala ADT preferable -- using compile-time specification or derivation rather than runtime reflection. This enables you to parse YAML into Json, and use your existing (or circe's generic) Decoders to perform the ADT marshalling. You can also use circe's Encoder to obtain a Json, and print that to YAML using this library.

The artifact is hosted by Sonatype, and release versions are synced to Maven Central:

For YAML 1.1

libraryDependencies += "io.circe" %% "circe-yaml" % "0.14.2"

or for YAML 1.2

libraryDependencies += "io.circe" %% "circe-yaml-v12" % "0.14.2"

Snapshot versions are available by adding the Sonatype Snapshots resolver:

resolvers ++= Resolver.sonatypeOssRepos("snapshots")

Parsing is accomplished through the io.circe.yaml.parser package; its API is similar to that of circe-parser:

import io.circe.yaml.parser
val json: Either[ParsingFailure, Json] = parser.parse(yamlString)

Additionally, there is a function for parsing multiple YAML documents from a single string:

val jsons: Stream[Either[ParsingFailure, Json]] = parser.parseDocuments(multiDocumentString)

Both of these methods also support a "streaming" parse from a java.io.Reader – this is different from the behavior of circe-streaming (which supports fully asynchronous streaming parsing with iteratees) but does provide a convenient way to retrieve YAML from Java inputs:

val config = getClass.getClassLoader.getResourceAsStream("config.yml")
val json = parser.parse(new InputStreamReader(config))

val configs = getClass.getClassLoader.getResourceAsStream("configs.yml")
val jsons = parser.parseDocuments(new InputStreamReader(configs))

Once you've parsed to Json, usage is the same as circe. For example, if you have circe-generic, you can do:

import cats.syntax.either._
import io.circe._
import io.circe.generic.auto._
import io.circe.yaml

case class Nested(one: String, two: BigDecimal)
case class Foo(foo: String, bar: Nested, baz: List[String])

val json = yaml.parser.parse("""
foo: Hello, World
bar:
    one: One Third
    two: 33.333333
baz:
    - Hello
    - World
""")

val foo = json
  .leftMap(err => err: Error)
  .flatMap(_.as[Foo])
  .valueOr(throw _)

Other features of YAML are supported:

• Multiple documents - use parseDocuments rather than parse to obtain Stream[Either[ParsingFailure, Json]]
• Streaming - use parse(reader: Reader) or parseDocuments(reader: Reader) to parse from a stream. Not sure what you'll get out of it.
• References / aliases - The reference will be replaced with the complete structure of the alias
• Explicit tags (on scalar values only) are handled by converting the tag/scalar pair into a singleton json object:

  example: !foo bar

  becomes

  { "example": { "foo": "bar" } }

The package io.circe.yaml.syntax provides an enrichment to Json which supports easily serializing to YAML using common options:

import cats.syntax.either._
import io.circe.yaml._
import io.circe.yaml.syntax._

val json = io.circe.jawn.parse("""{"foo":"bar"}""").valueOr(throw _)

println(json.asYaml.spaces2) // 2 spaces for each indent level
println(json.asYaml.spaces4) // 4 spaces for each indent level

Additionally, there is a class io.circe.yaml.Printer which (in similar fashion to circe's Printer) can be configured with many options which control the String output. Its pretty method produces a String using the configured options:

io.circe.yaml.Printer(dropNullKeys = true, mappingStyle = Printer.FlowStyle.Block)
  .pretty(json)

Only JSON-compatible YAML can be used, for obvious reasons:

• Complex keys are not supported (only String keys)
• Unlike YAML collections, a JSON array is not the same as a JSON object with integral keys (given the above, it would be impossible). So, a YAML mapping with integral keys will still be a JSON object, and the keys will be strings.

This is released under the Apache 2.0 license, as specified in the LICENSE file. It depends on both circe and SnakeYAML, which each has its own license. Consult those projects to learn about their licenses.

This library is neither endorsed by, nor affiliated with, SnakeYAML.

As part of the circe community, circe-yaml supports the Typelevel code of conduct and wants all of its channels (Gitter, GitHub, etc.) to be welcoming environments for everyone.

Please read the circe Contributor's Guide for information about how to submit a pull request.

This circe community module is currently maintained by Jeff May, Darren Gibson, and Zach McCoy. It strives to conform as closely as possible to the style of circe itself.