JSON Schema validator. Supports drafts 4, 6, 7, 2019-09, 2020-12, OpenAPI 3.0, and OpenAPI 3.1.
Add this line to your application's Gemfile:
gem 'json_schemer'
And then execute:
$ bundle
Or install it yourself as:
$ gem install json_schemer
require 'json_schemer'
schema = {
'type' => 'object',
'properties' => {
'abc' => {
'type' => 'integer',
'minimum' => 11
}
}
}
schemer = JSONSchemer.schema(schema)
# true/false validation
schemer.valid?({ 'abc' => 11 })
# => true
schemer.valid?({ 'abc' => 10 })
# => false
# error validation (`validate` returns an enumerator)
schemer.validate({ 'abc' => 10 }).to_a
# => [{"data"=>10,
# "data_pointer"=>"/abc",
# "schema"=>{"type"=>"integer", "minimum"=>11},
# "schema_pointer"=>"/properties/abc",
# "root_schema"=>{"type"=>"object", "properties"=>{"abc"=>{"type"=>"integer", "minimum"=>11}}},
# "type"=>"minimum",
# "error"=>"number at `/abc` is less than: 11"}]
# default property values
data = {}
JSONSchemer.schema(
{
'properties' => {
'foo' => {
'default' => 'bar'
}
}
},
insert_property_defaults: true
).valid?(data)
data
# => {"foo"=>"bar"}
# schema files
require 'pathname'
schema = Pathname.new('/path/to/schema.json')
schemer = JSONSchemer.schema(schema)
# schema json string
schema = '{ "type": "integer" }'
schemer = JSONSchemer.schema(schema)
# schema validation
JSONSchemer.valid_schema?({ '$id' => 'valid' })
# => true
JSONSchemer.validate_schema({ '$id' => '#invalid' }).to_a
# => [{"data"=>"#invalid",
# "data_pointer"=>"/$id",
# "schema"=>{"$ref"=>"#/$defs/uriReferenceString", "$comment"=>"Non-empty fragments not allowed.", "pattern"=>"^[^#]*#?$"},
# "schema_pointer"=>"/properties/$id",
# "root_schema"=>{...meta schema},
# "type"=>"pattern",
# "error"=>"string at `/$id` does not match pattern: ^[^#]*#?$"}]
# subschemas
schema = {
'type' => 'integer',
'$defs' => {
'foo' => {
'type' => 'string'
}
}
}
schemer = JSONSchemer.schema(schema)
schemer.ref('#/$defs/foo').validate(1).to_a
# => [{"data"=>1,
# "data_pointer"=>"",
# "schema"=>{"type"=>"string"},
# "schema_pointer"=>"/$defs/foo",
# "root_schema"=>{"type"=>"integer", "$defs"=>{"foo"=>{"type"=>"string"}}},
# "type"=>"string",
# "error"=>"value at root is not a string"}]
# schema bundling (https://json-schema.org/draft/2020-12/json-schema-core.html#section-9.3)
schema = {
'$id' => 'http://example.com/schema',
'allOf' => [
{ '$ref' => 'schema/one' },
{ '$ref' => 'schema/two' }
]
}
refs = {
URI('http://example.com/schema/one') => {
'type' => 'integer'
},
URI('http://example.com/schema/two') => {
'minimum' => 11
}
}
schemer = JSONSchemer.schema(schema, :ref_resolver => refs.to_proc)
schemer.bundle
# => {"$id"=>"http://example.com/schema",
# "allOf"=>[{"$ref"=>"schema/one"}, {"$ref"=>"schema/two"}],
# "$schema"=>"https://json-schema.org/draft/2020-12/schema",
# "$defs"=>
# {"http://example.com/schema/one"=>{"type"=>"integer", "$id"=>"http://example.com/schema/one", "$schema"=>"https://json-schema.org/draft/2020-12/schema"},
# "http://example.com/schema/two"=>{"minimum"=>11, "$id"=>"http://example.com/schema/two", "$schema"=>"https://json-schema.org/draft/2020-12/schema"}}}
JSONSchemer.schema(
schema,
# meta schema to use for vocabularies (keyword behavior) and schema validation
# String/JSONSchemer::Schema
# 'https://json-schema.org/draft/2020-12/schema': JSONSchemer.draft202012
# 'https://json-schema.org/draft/2019-09/schema': JSONSchemer.draft201909
# 'http://json-schema.org/draft-07/schema#': JSONSchemer.draft7
# 'http://json-schema.org/draft-06/schema#': JSONSchemer.draft6
# 'http://json-schema.org/draft-04/schema#': JSONSchemer.draft4
# 'http://json-schema.org/schema#': JSONSchemer.draft4
# 'https://spec.openapis.org/oas/3.1/dialect/base': JSONSchemer.openapi31
# 'json-schemer://openapi30/schema': JSONSchemer.openapi30
# default: JSONSchemer.draft202012
meta_schema: 'https://json-schema.org/draft/2020-12/schema',
# validate `format` (https://json-schema.org/draft/2020-12/json-schema-validation.html#section-7)
# true/false
# default: true
format: true,
# custom formats
formats: {
'int32' => proc do |instance, _format|
instance.is_a?(Integer) && instance.bit_length <= 32
end,
# disable specific format
'email' => false
},
# custom content encodings
# only `base64` is available by default
content_encodings: {
# return [success, annotation] tuple
'urlsafe_base64' => proc do |instance|
[true, Base64.urlsafe_decode64(instance)]
rescue
[false, nil]
end
},
# custom content media types
# only `application/json` is available by default
content_media_types: {
# return [success, annotation] tuple
'text/csv' => proc do |instance|
[true, CSV.parse(instance)]
rescue
[false, nil]
end
},
# insert default property values during validation
# string keys by default (use `:symbol` to insert symbol keys)
# true/false/:symbol
# default: false
insert_property_defaults: true,
# modify properties during validation. You can pass one Proc or a list of Procs to modify data.
# Proc/[Proc]
# default: nil
before_property_validation: proc do |data, property, property_schema, _parent|
data[property] ||= 42
end,
# modify properties after validation. You can pass one Proc or a list of Procs to modify data.
# Proc/[Proc]
# default: nil
after_property_validation: proc do |data, property, property_schema, _parent|
data[property] = Date.iso8601(data[property]) if property_schema.is_a?(Hash) && property_schema['format'] == 'date'
end,
# resolve external references
# 'net/http'/proc/lambda/respond_to?(:call)
# 'net/http': proc { |uri| JSON.parse(Net::HTTP.get(uri)) }
# default: proc { |uri| raise UnknownRef, uri.to_s }
ref_resolver: 'net/http',
# use different method to match regexes
# 'ruby'/'ecma'/proc/lambda/respond_to?(:call)
# 'ruby': proc { |pattern| Regexp.new(pattern) }
# default: 'ruby'
regexp_resolver: proc do |pattern|
RE2::Regexp.new(pattern)
end,
# output formatting (https://json-schema.org/draft/2020-12/json-schema-core.html#section-12)
# 'classic'/'flag'/'basic'/'detailed'/'verbose'
# default: 'classic'
output_format: 'basic',
# validate `readOnly`/`writeOnly` keywords (https://spec.openapis.org/oas/v3.0.3#fixed-fields-19)
# 'read'/'write'/nil
# default: nil
access_mode: 'read'
)
Configuration options can be set globally by modifying JSONSchemer.configuration
. Global options are applied to any new schemas at creation time (global configuration changes are not reflected in existing schemas). They can be overridden with the regular keyword arguments described above.
# configuration block
JSONSchemer.configure do |config|
config.regexp_resolver = 'ecma'
end
# configuration accessors
JSONSchemer.configuration.insert_property_defaults = true
Error messages can be customized using the x-error
keyword and/or I18n translations. x-error
takes precedence if both are defined.
# override all errors for a schema
schemer = JSONSchemer.schema({
'type' => 'string',
'x-error' => 'custom error for schema and all keywords'
})
schemer.validate(1).first
# => {"data"=>1,
# "data_pointer"=>"",
# "schema"=>{"type"=>"string", "x-error"=>"custom error for schema and all keywords"},
# "schema_pointer"=>"",
# "root_schema"=>{"type"=>"string", "x-error"=>"custom error for schema and all keywords"},
# "type"=>"string",
# "error"=>"custom error for schema and all keywords",
# "x-error"=>true}
schemer.validate(1, :output_format => 'basic')
# => {"valid"=>false,
# "keywordLocation"=>"",
# "absoluteKeywordLocation"=>"json-schemer://schema#",
# "instanceLocation"=>"",
# "error"=>"custom error for schema and all keywords",
# "x-error"=>true,
# "errors"=>#<Enumerator: ...>}
# keyword-specific errors
schemer = JSONSchemer.schema({
'type' => 'string',
'minLength' => 10,
'x-error' => {
'type' => 'custom error for `type` keyword',
# special `^` keyword for schema-level error
'^' => 'custom error for schema',
# same behavior as when `x-error` is a string
'*' => 'fallback error for schema and all keywords'
}
})
schemer.validate(1).map { _1.fetch('error') }
# => ["custom error for `type` keyword"]
schemer.validate('1').map { _1.fetch('error') }
# => ["custom error for schema and all keywords"]
schemer.validate(1, :output_format => 'basic').fetch('error')
# => "custom error for schema"
# variable interpolation (instance/instanceLocation/keywordLocation/absoluteKeywordLocation)
schemer = JSONSchemer.schema({
'$id' => 'https://example.com/schema',
'properties' => {
'abc' => {
'type' => 'string',
'x-error' => <<~ERROR
instance: %{instance}
instance location: %{instanceLocation}
keyword location: %{keywordLocation}
absolute keyword location: %{absoluteKeywordLocation}
ERROR
}
}
})
puts schemer.validate({ 'abc' => 1 }).first.fetch('error')
# instance: 1
# instance location: /abc
# keyword location: /properties/abc/type
# absolute keyword location: https://example.com/schema#/properties/abc/type
When the I18n gem is loaded, custom error messages are looked up under the json_schemer
key. It may be necessary to restart your application after adding the root key because the existence check is cached for performance reasons.
Translation keys are looked up in this order:
$LOCALE.json_schemer.errors.$ABSOLUTE_KEYWORD_LOCATION
$LOCALE.json_schemer.errors.$SCHEMA_ID.$KEYWORD_LOCATION
$LOCALE.json_schemer.errors.$KEYWORD_LOCATION
$LOCALE.json_schemer.errors.$SCHEMA_ID.$KEYWORD
$LOCALE.json_schemer.errors.$SCHEMA_ID.*
$LOCALE.json_schemer.errors.$META_SCHEMA_ID.$KEYWORD
$LOCALE.json_schemer.errors.$META_SCHEMA_ID.*
$LOCALE.json_schemer.errors.$KEYWORD
$LOCALE.json_schemer.errors.*
Example translations file:
en:
json_schemer:
errors:
'https://example.com/schema#/properties/abc/type': custom error for absolute keyword location
'https://example.com/schema':
'#/properties/abc/type': custom error for keyword location, nested under schema $id
'type': custom error for `type` keyword, nested under schema $id
'^': custom error for schema, nested under schema $id
'*': fallback error for schema and all keywords, nested under schema $id
'#/properties/abc/type': custom error for keyword location
'http://json-schema.org/draft-07/schema#':
'type': custom error for `type` keyword, nested under meta-schema $id ($schema)
'^': custom error for schema, nested under meta-schema $id
'*': fallback error for schema and all keywords, nested under meta-schema $id ($schema)
'type': custom error for `type` keyword
'^': custom error for schema
# variable interpolation (instance/instanceLocation/keywordLocation/absoluteKeywordLocation)
'*': |
fallback error for schema and all keywords
instance: %{instance}
instance location: %{instanceLocation}
keyword location: %{keywordLocation}
absolute keyword location: %{absoluteKeywordLocation}
And output:
require 'i18n'
I18n.locale = :en # $LOCALE=en
schemer = JSONSchemer.schema({
'$id' => 'https://example.com/schema', # $SCHEMA_ID=https://example.com/schema
'$schema' => 'http://json-schema.org/draft-07/schema#', # $META_SCHEMA_ID=http://json-schema.org/draft-07/schema#
'properties' => {
'abc' => {
'type' => 'integer' # $KEYWORD=type
} # $KEYWORD_LOCATION=#/properties/abc/type
} # $ABSOLUTE_KEYWORD_LOCATION=https://example.com/schema#/properties/abc/type
})
schemer.validate({ 'abc' => 'not-an-integer' }).first
# => {"data"=>"not-an-integer",
# "data_pointer"=>"/abc",
# "schema"=>{"type"=>"integer"},
# "schema_pointer"=>"/properties/abc",
# "root_schema"=>{"$id"=>"https://example.com/schema", "$schema"=>"http://json-schema.org/draft-07/schema#", "properties"=>{"abc"=>{"type"=>"integer"}}},
# "type"=>"integer",
# "error"=>"custom error for absolute keyword location",
# "i18n"=>true
In the example above, custom error messsages are looked up using the following keys (in order until one is found):
en.json_schemer.errors.'https://example.com/schema#/properties/abc/type'
en.json_schemer.errors.'https://example.com/schema'.'#/properties/abc/type'
en.json_schemer.errors.'#/properties/abc/type'
en.json_schemer.errors.'https://example.com/schema'.type
en.json_schemer.errors.'https://example.com/schema'.*
en.json_schemer.errors.'http://json-schema.org/draft-07/schema#'.type
en.json_schemer.errors.'http://json-schema.org/draft-07/schema#'.*
en.json_schemer.errors.type
en.json_schemer.errors.*
document = JSONSchemer.openapi({
'openapi' => '3.1.0',
'info' => {
'title' => 'example'
},
'components' => {
'schemas' => {
'example' => {
'type' => 'integer'
}
}
}
})
# document validation using meta schema
document.valid?
# => false
document.validate.to_a
# => [{"data"=>{"title"=>"example"},
# "data_pointer"=>"/info",
# "schema"=>{...info schema},
# "schema_pointer"=>"/$defs/info",
# "root_schema"=>{...meta schema},
# "type"=>"required",
# "details"=>{"missing_keys"=>["version"]}},
# ...]
# data validation using schema by name (in `components/schemas`)
document.schema('example').valid?(1)
# => true
document.schema('example').valid?('one')
# => false
# data validation using schema by ref
document.ref('#/components/schemas/example').valid?(1)
# => true
document.ref('#/components/schemas/example').valid?('one')
# => false
The json_schemer
executable takes a JSON schema file as the first argument followed by one or more JSON data files to validate. If there are any validation errors, it outputs them and returns an error code.
Validation errors are output as single-line JSON objects. The --errors
option can be used to limit the number of errors returned or prevent output entirely (and fail fast).
The schema or data can also be read from stdin using -
.
% json_schemer --help
Usage:
json_schemer [options] <schema> <data>...
json_schemer [options] <schema> -
json_schemer [options] - <data>...
json_schemer -h | --help
json_schemer --version
Options:
-e, --errors MAX Maximum number of errors to output
Use "0" to validate with no output
-h, --help Show help
-v, --version Show version
After checking out the repo, run bin/setup
to install dependencies. Then, run rake test
to run the tests. You can also run bin/console
for an interactive prompt that will allow you to experiment.
Bug reports and pull requests are welcome on GitHub at https://github.com/davishmcclurg/json_schemer.
The gem is available as open source under the terms of the MIT License.