Class: Fog::Schema::DataValidator

Inherits:

Object

• Object
• Fog::Schema::DataValidator

Defined in:

lib/fog/schema/data_validator.rb

Overview

This validates a data object against a Ruby based schema to see if they match.

• An object matches the schema if == or === returns true

• Hashes match if all the key’s values match the classes given in the schema as well. This can be configured in the options

• Arrays match when every element in the data matches the case given in the schema.

The schema and validation are very simple and probably not suitable for some cases.

The following classes can be used to check for special behaviour

• Fog::Boolean - value may be true or false

• Fog::Nullable::Boolean - value may be true, false or nil

• Fog::Nullable::Integer - value may be an Integer or nil

• Fog::Nullable::String

• Fog::Nullable::Time

• Fog::Nullable::Float

• Fog::Nullable::Hash

• Fog::Nullable::Array

All the “nullable” objects will pass if the value is of the class or if it is nil. This allows you to match APIs that may include keys when the value is not available in some cases but will always be a String. Such as an password that is only displayed on the reset action.

The keys for “nullable” resources should always be present but original matcher had a bug that allowed these to also appear to work as optional keys/values.

If you need the original behaviour, data with a missing key is still valid, then you may pass the :allow_optional_rules option to the #validate method.

That is not recommended because you are describing a schema with optional keys in a format that does not support it.

Setting :allow_extra_keys as true allows the data to contain keys not declared by the schema and still pass. This is useful if new attributes appear in the API in a backwards compatible manner and can be ignored.

This is the behaviour you would have seen with strict being false in the original test helper.

Examples:

Schema example

{
  "id" => String,
  "ram" => Integer,
  "disks" => [
    "size" => Float
  ],
  "dns_name" => Fog::Nullable::String,
  "active" => Fog::Boolean,
  "created" => DateTime
}

Instance Attribute Summary

• #message ⇒ String readonly

  This returns the last message set by the validator.

Instance Method Summary

• #initialize ⇒ DataValidator constructor

  A new instance of DataValidator.

• #validate(data, schema, options = {}) ⇒ Boolean

  Checks if the data structure matches the schema passed in and returns true if it fits.

Constructor Details

#initialize ⇒ DataValidator

Returns a new instance of DataValidator.

# File 'lib/fog/schema/data_validator.rb', line 69

def initialize
  @message = nil
end

Instance Attribute Details

#message ⇒ String (readonly)

This returns the last message set by the validator

Returns:

• (String)

# File 'lib/fog/schema/data_validator.rb', line 67

def message
  @message
end

Instance Method Details

#validate(data, schema, options = {}) ⇒ Boolean

Checks if the data structure matches the schema passed in and returns true if it fits.

Parameters:

• data (Object) —

  Hash or Array to check

• schema (Object) —

  Schema pattern to check against

• options (Boolean) (defaults to: {})

Options Hash (options):

• :allow_extra_keys (Boolean) —

  If true does not fail if extra keys are in the data that are not in the schema.

• :allow_optional_rules (Boolean) —

  If true does not fail if extra keys are in the schema that do not match the data. Not recommended!

Returns:

• (Boolean) —

  Did the data fit the schema?

# File 'lib/fog/schema/data_validator.rb', line 87

def validate(data, schema, options = {})
  valid = validate_value(schema, data, options)

  unless valid
    @message = "#{data.inspect} does not match #{schema.inspect}"
  end
  valid
end