Constraints

Schema defines several places for validation constraints that may be used to analyze user input. List of such constraints is not limited, while this schema uses a finite set of constraints, custom schema extensions may add their own constraints, so schema peers should not break on unknown constraints. It is also considered to be good form to prefix custom constraint names to prevent name clashing.

Following are the constraints used throughout default schema:

notNull

Validates that target is not null

type: notNull

boolean

Validates that target is a boolean value

type: boolean

string

Validates that target is a string

type: string
minimumLength: 10
maximumLength: 12

number

Validates that target is numeric (any integer or real number)

type: number
minimum: 0
maximum: 10.5
exclusiveMaximum: true
exclusiveMinimum: false

integer

Validates that target is an integer. Size (short, int, long, etc) is not taken into account.

type: integer
minimum: 0
maximum: 10
exclusiveMaximum: true
exclusiveMinimum: false

real

Validates that target is a floating-point arithmetic number. Integer numbers must pass this validation if implementation language coerces numbers like 1.0 to 1. The precision (float, double, etc.) is not taken into account.

type: real
minimum: 0.5
maximum: 10.5
exclusiveMaximum: true
exclusiveMinimum: false

array

Validates that target is a list

type: array

object

Validates that target is a key-value map / hash

type: object

amount

Special ElasticSearch constraint for validating values like 10, 2k, 10m

type: amount

time

Special ElasticSearch constraint for validating values like 10s, 2m, 10h

type: time

anyOf

Validates that at least one of specified constraints is satisfied by target

type: anyOf
constraints:
  - type: string
  - type: boolean

values

Applies specified constraints to all array/object values

type: values
constraints:
  - type: string
  - type: notNull

nested

Contains specific constraints for object properties

type: nested
properties:
  aspects:
    - type: object
    - type: values
      constraints:
        - type: string
  mapping:
    - type: mappingType

size

Validates target size

type: size
maximum: 12
minimum: 11

mappingType

Ensures that target specifies a valid mapping type. Constraint validator should refer to current schema to validate such value.

type: mappingType

glob

Ensures that target matches specified glob pattern:

type: glob
pattern: '*=>*'

Implicit constraints

Beside explicit constraints, there are three types of implicit constraints. First, if the entity hasn’t been marked as nullable: true, a nonNull constraint should be added:

type: string

Should be treated as

type: string
constraints:
  - type: nonNull

Second, every specified type should be treated like constraint of same name:

type: string
nullable: true

implies:

constraints:
  - type: string

If there is list of types specified, value should match any of those.

Finally, there is a case with predefined set of values:

type: string
values:
  - alpha
  - beta
  - gamma

Which should be treated as something like

constraints:
  - type: anyValue
    configuration:
      values:
        - alpha
        - beta
        - gamma

Enum validation

Validator service should be aware of enums. Such values may be wrapped in array, so validator should apply all constraints either to root or every element.