Enum Schema

pub enum Schema {
    Empty {
        definitions: Definitions,
        metadata: Metadata,
    },
    Ref {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        ref_: String,
    },
    Type {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        type_: Type,
    },
    Enum {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        enum_: BTreeSet<String>,
    },
    Elements {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        elements: Box<Schema>,
    },
    Properties {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        properties: BTreeMap<String, Schema>,
        optional_properties: BTreeMap<String, Schema>,
        properties_is_present: bool,
        additional_properties: bool,
    },
    Values {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        values: Box<Schema>,
    },
    Discriminator {
        definitions: Definitions,
        metadata: Metadata,
        nullable: bool,
        discriminator: String,
        mapping: BTreeMap<String, Schema>,
    },
}

A pattern-matching-friendly representation of a JSON Typedef schema.

Each variant of this schema corresponds to one of the eight “forms” a schema may take on. All of the forms share the following fields:

• definitions corresponds to the JSON Typedef keyword of the same name. This should only be non-empty on root schemas. Otherwise, Schema::validate will return SchemaValidateError::NonRootDefinitions.

• metadata corresponds to the JSON Typedef keyword of the same name. Use this to convey information not pertinent to validation, such as hints for code generation. Do not expect other parties to understand the fields inside metadata unless you’ve agreed upon them out-of-band.

Except for Schema::Empty, all of the forms also share one additional field:

• nullable corresponds to the JSON Typedef keyword of the same name. If set to “true”, then regardless of any other considerations the schema will accept JSON null as valid.

Schema::Empty omits nullable because it’s redundant; schemas of the empty form already accept null anyway.

For convenience, these three common properties have associated borrowing “getters”: Schema::definitions, Schema::metadata, and Schema::nullable.

If you are trying to parse a JSON Typedef schema from JSON, see SerdeSchema and Schema::from_serde_schema.

use jtd::{SerdeSchema, Schema};
use serde_json::json;

assert_eq!(
    Schema::from_serde_schema(serde_json::from_value(json!({
        "elements": {
            "type": "uint32",
            "nullable": true
        }
    })).unwrap()).unwrap(),
    jtd::Schema::Elements {
        definitions: Default::default(),
        metadata: Default::default(),
        nullable: false,
        elements: Box::new(jtd::Schema::Type {
            definitions: Default::default(),
            metadata: Default::default(),
            nullable: true,
            type_: jtd::Type::Uint32,
        })
    }
);

Variants

Empty

The empty form.

The empty form will accept all inputs. It corresponds to the “top” type of many programming language, like Java’s Object or TypeScript’s any.

Fields

definitions: Definitions

metadata: Metadata

Ref

The ref form.

The ref form accepts whatever the definition it refers to accepts.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

ref_: String

The name of the definition being referred to.

Type

The type form.

The type form accepts JSON “primitives” (booleans, numbers, strings) whose value fall within a certain “type”. These types are enumerated in Type.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

type_: Type

The type of primitive value accepted.

Enum

The enum form.

The enum form accepts JSON strings whose values are within an enumerated set.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

enum_: BTreeSet<String>

The values the schema accepts.

Elements

The elements form.

The elements form accepts JSON arrays, and each element of the array is validated against a sub-schema.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

elements: Box<Schema>

A schema for the elements of the array.

Properties

The properties form.

The properties form accepts JSON objects being used as “structs”.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

properties: BTreeMap<String, Schema>

The required properties of the “struct”, and the schema that each must satisfy.

optional_properties: BTreeMap<String, Schema>

The optional properties of the “struct”, and the schema that each must satisfy if present.

properties_is_present: bool

Whether the properties keyword is present on the schema.

It is invalid to set this to false while having properties be non-empty.

This is used only to handle the corner case of a properties-form schema being used to validate a non-object; in order to ensure the returned schema_path points to a part of the schema that really exists, validators need to be able to tell the difference between properties being an empty object versus being omitted from the schema.

This field does not affect whether an input is valid. It only affects the schema_path that will be returned if that input is not an object. For more details, see the first sub-bullet after “Otherwise” in RFC 8927, Section 3.3.6.

Schema::from_serde_schema correctly handles populating this field. If you are constructing schemas by hand and want to play it safe, it is always safe to set this to true.

additional_properties: bool

Whether additional properties not specified in properties or optional_properties are permitted.

Values

The values form.

The values form accepts JSON objects being used as “dictionaries”; each value of the dictionary is validated against a sub-schema.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

values: Box<Schema>

A schema for the values of the “dictionary” object.

Discriminator

The discriminator form.

The discriminator form accepts JSON objects being used as “discriminated unions”, or “tagged unions”.

Fields

definitions: Definitions

metadata: Metadata

nullable: bool

discriminator: String

The “discriminator” property of the schema.

For an input to be valid, this property must exist and its value must be a key in mapping.

mapping: BTreeMap<String, Schema>

A mapping from the value of the discriminator property in the input to a schema that the rest of the input (without the discriminator property) must satisfy.

Implementations

impl Schema

pub fn into_serde_schema(self) -> SerdeSchema

Converts a Schema into a SerdeSchema.

use jtd::{Schema, SerdeSchema, Type};

assert_eq!(
    SerdeSchema {
        type_: Some("uint8".to_owned()),
        ..Default::default()
    },
    Schema::Type {
        definitions: Default::default(),
        metadata: Default::default(),
        nullable: false,
        type_: Type::Uint8,
    }.into_serde_schema(),
);

pub fn from_serde_schema( serde_schema: SerdeSchema, ) -> Result<Self, FromSerdeSchemaError>

Constructs a Schema from a SerdeSchema.

use jtd::{Schema, SerdeSchema, Type};

assert_eq!(
    Schema::Type {
        definitions: Default::default(),
        metadata: Default::default(),
        nullable: false,
        type_: Type::Uint8,
    },
    Schema::from_serde_schema(SerdeSchema {
        type_: Some("uint8".to_owned()),
        ..Default::default()
    }).unwrap(),
);

See the documentation for FromSerdeSchemaError for examples of how this function may return an error.

pub fn validate(&self) -> Result<(), SchemaValidateError>

Ensures a Schema is well-formed.

use jtd::{Schema, Type};

let schema = Schema::Type {
    definitions: Default::default(),
    metadata: Default::default(),
    nullable: false,
    type_: Type::Uint8,
};

schema.validate().expect("Invalid schema");

See the documentation for SchemaValidateError for examples of how this function may return an error.

pub fn definitions(&self) -> &BTreeMap<String, Schema>

Gets the schema’s definitions.

use jtd::{Definitions, Schema};

assert_eq!(
    &vec![(
        "foo".to_owned(),
        Schema::Empty {
            definitions: Default::default(),
            metadata: Default::default(),
        },
    )].into_iter().collect::<Definitions>(),

     Schema::Empty {
         definitions: vec![(
            "foo".to_owned(),
            Schema::Empty {
                definitions: Default::default(),
                metadata: Default::default(),
            },
        )].into_iter().collect(),
         metadata: Default::default(),
     }.definitions(),
);

pub fn metadata(&self) -> &BTreeMap<String, Value>

Gets the schema’s metadata.

use jtd::{Metadata, Schema};
use serde_json::json;

assert_eq!(
    &vec![(
        "foo".to_owned(),
        json!("bar"),
    )].into_iter().collect::<Metadata>(),

    Schema::Empty {
        definitions: Default::default(),
        metadata: vec![(
           "foo".to_owned(),
           json!("bar"),
       )].into_iter().collect(),
    }.metadata(),
);

pub fn nullable(&self) -> bool

Gets whether the schema is nullable.

For Schema::Empty, this always returns true. For all other forms, this fetches the nullable property.

use jtd::{Schema, Type};

assert!(
    Schema::Empty {
        definitions: Default::default(),
        metadata: Default::default(),
    }.nullable(),
);

assert!(
    !Schema::Type {
        definitions: Default::default(),
        metadata: Default::default(),
        nullable: false,
        type_: Type::Uint8,
    }.nullable(),
);

Trait Implementations

impl Clone for Schema

fn clone(&self) -> Schema

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Schema

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for Schema

fn eq(&self, other: &Schema) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for Schema