$schema: "http://json-schema.org/schema#"
title: "GA4GH-VR-Definitions"
type: object

definitions:
  # VR definitions are presented top-down.  Everything rolls up to
  # Variation, which is a polymorphic abstraction of many kinds of
  # variation.


  ############################################################################
  # Variation

  Variation:
    additionalProperties: false
    description: >-
      A polymorphic representation of variation
    oneOf:
      - $ref: "#/definitions/Text"
      - $ref: "#/definitions/Allele"
      # future expansion of variation types will go here
    discriminator:
      propertyName: type

  Text:
    additionalProperties: false
    type: object
    properties:
      type:
        type: string
        enum: [Text]
        default: Text
      _id:
        $ref: "#/definitions/CURIE"
      definition:
        type: string
        description: >-
          An textual representation of variation intended to capture
          variation descriptions that cannot be parsed, but still
          treated as variation.

  Allele:
    additionalProperties: false
    description: An assertion of a sequence state at a Location.
    type: object
    properties:
      type:
        type: string
        enum: [Allele]
        default: Allele
      _id:
        $ref: "#/definitions/CURIE"
      location:
        $ref: "#/definitions/Location"
      state:
        $ref: "#/definitions/State"


  ############################################################################
  # Locations

  Location:
    additionalProperties: false
    description: >-
      A Location represents a span on a specific sequence.
    oneOf:
      - $ref: "#/definitions/SequenceLocation"
    discriminator:
      propertyName: type

  SequenceLocation:
    additionalProperties: false
    type: object
    properties:
      type:
        type: string
        enum: [SequenceLocation]
        default: SequenceLocation
      _id:
        $ref: "#/definitions/CURIE"
      sequence_id:
        $ref: "#/definitions/CURIE"
      interval:
        $ref: "#/definitions/Interval"


  ############################################################################
  # Interval
  # * SimpleInterval

  Interval:
    additionalProperties: false
    description: >-
      An Interval represents a span of sequence. Positions are
      always represented by contiguous spans using interbase
      coordinates.

      This definition of Interval is intended to be compatible with that
      in Sequence Ontology
      ([SO:0000001](http://www.sequenceontology.org/browser/current_svn/term/SO:0000001)),
      with the exception that the GA4GH VR Interval may be
      zero-width. The SO definition is for an "extent greater than
      zero".
    oneOf:
      - $ref: '#/definitions/SimpleInterval'
    discriminator:
      propertyName: type

  SimpleInterval:
    additionalProperties: false
    type: object
    properties:
      type:
        type: string
        enum: [SimpleInterval]
        default: SimpleInterval
      start:
        type: integer
        nullable: true
      end:
        type: integer
        nullable: true
    required:
      - start
      - end
    example:
      type: SimpleInterval
      start: 11
      end: 22


  ############################################################################
  # States

  State:
    oneOf:
      - $ref: "#/definitions/SequenceState"

  SequenceState:
    additionalProperties: false
    type: object
    properties:
      type:
        type: string
        enum: [SequenceState]
        default: SequenceState
      sequence:
        type: string
        nullable: false
    required:
      - type
      - sequence
    example:
      type: SequenceState
      sequence: C


  ############################################################################
  # Basic types
  # These types are used solely within other definitions.

  DateTime:
    additionalProperties: false
    type: string
    format: "date-time"

  CURIE:
    additionalProperties: false
    description: >-
      A string that refers to an object uniquely.  The lifetime and
      scope of an id is defined by the sender.

      VR does not impose any contraints on strings used as ids in
      messages. However, to maximize sharability of data, the VR
      Specification RECOMMENDS that implementations use [W3C Compact
      URI (CURIE)](https://www.w3.org/TR/curie/) syntax.

      String CURIEs are represented as `prefix`:`reference` (W3C
      terminology), but often referred to as `namespace`:`accession`
      or `namespace`:`local id` colloquially.

      The VR specification also RECOMMENDS that `prefix` be defined in
      identifiers.org.

      The `reference` component is an unconstrained string.

      A CURIE is a URI.  URIs may *locate* objects (i.e., specify
      where to retrieve them) or *name* objects conceptually.  VR uses
      CURIEs primarily as a naming mechanism.

      Implementations MAY provide CURIE resolution mechanisms for
      prefixes to make these objects locatable.

      Using internal ids in public messages is strongly discouraged.

    type: string
    pattern: '^\w[^:]+:.+$'
    example: "ga4gh:GA.01234abcde"