Data.JsonSchema

Primary API

compile :: Spec -> Graph -> RawSchema -> Schema

validate :: Schema -> Value -> Vector ValErr

Types

newtype Spec

Constructors

Spec

Fields

_unSpec :: HashMap Text (ValidatorGen, EmbeddedSchemas)

type Graph = HashMap Text (HashMap Text Value)

Set of potentially mutually recursive schemas.

type ValErr = Text

type Validator = Value -> Vector ValErr

type Schema = Vector Validator

type ValidatorGen = Spec -> Graph -> RawSchema -> Value -> Maybe Validator

type EmbeddedSchemas = Text -> Value -> Vector RawSchema

Return a schema's immediate subschemas.

This is used by fetchRefs to find all the subschemas in a document. This allows it to process only "$ref"s and "id"s that are actual schema keywords.

data RawSchema

Constructors

RawSchema

Fields

_rsURI :: Text

_rsObject :: HashMap Text Value

Draft 4 Specific

draft4 :: Spec

isValidSchema :: RawSchema -> Vector ValErr

Check if a RawSchema is valid Draft 4 schema.

This is just a convenience function built by preloading validate with the spec schema that describes valid Draft 4 schemas.

NOTE: It's not actually required to run isValidSchema on prospective draft 4 schemas at all. However, it's a good way to catch unintentional mistakes in schema documents.

compileDraft4 :: Graph -> RawSchema -> Either (Vector ValErr) Schema

Check that a RawSchema conforms to the JSON Schema Draft 4 master schema document. Compile it if it does.

This is just a convenience function built by combining isValidSchema and compile.

Graph Builder

TODO: It would be nice to have a few other functions for building out Graphs depending on the situation. For instance, there could be a function that only fetched "$ref"s which use the "file://" scheme.

fetchRefs :: Spec -> RawSchema -> Graph -> IO (Either Text Graph)

Take a schema. Retrieve every document either it or its subschemas include via the "$ref" keyword. Load a Graph out with them.

TODO: This function's URL processing is hacky and needs improvement.