data-validator

This is a Ruby version of gfx/p5-Data-Validator + xaicron/p5-Data-Validator-Recursive.

Compare below versus https://metacpan.org/pod/Data::Validator

rule = Data::Validator.new(
  'uri'        => { isa: String, xor: %w(schema host path_query) },
  'schema'     => { isa: String, default: 'http' },
  'host'       => { isa: String },
  'path_query' => { isa: String, default: '/' },
  'method'     => { isa: String, default: 'GET' },
)

args = rule.validate('uri' => 'http://example.com')

and below versus https://metacpan.org/pod/Data::Validator::Recursive

# create a new rule
rule = Data::Validator::Recursive.new(
    'foo' => String,
    'bar' => { isa: Integer },
    'baz' => {
        isa: Hash, # default
        rule: {
            'hoge' => { isa: String, optional: 1 },
            'fuga' => Integer
        },
    },
)

# input data for validation
input = {
    'foo' => 'hoge',
    'bar' => 1192,
    'baz' => {
        'hoge' => 'kamakura',
        'fuga' => 1185,
    },
}

# do validation
params = rule.validate(input) # raises automatically on error

limitations

• Data::Validator is recursive by default. There is no such thing like a nonrecursive validator.
• ->with('Method') does not make sense to us, so not supported.
• does also does not make sense. Not supported.
• We do distinguish arrays and hashes unlike perl. There also are no ->with('Sequenced').
• I don't understand the actual needs of xor; all examples seems illustravive to me. Other validators like JSON Schema (cf zigorou/perl-JSV) do not have this. This lack of understanding can negatively impact.
• I don't understand why @gfx thinks it's fast.