Validation Library Support
There are official adapters available for zod
and yup
,
but you can easily support whatever library you want by creating your own adapter.
And if you create an adapter for a library, feel free to make a PR to add official support 😊
Creating an adapter
Any object that conforms to the Validator
type can be passed into the the ValidatedForm
's validator
prop.
type FieldErrors = Record<string, string>;
type ValidationResult<DataType> =
| { data: DataType; error: undefined }
| { error: FieldErrors; data: undefined };
type ValidateFieldResult = { error?: string };
type Validator<DataType> = {
validate: (
unvalidatedData: unknown,
) => Promise<ValidationResult<DataType>>;
validateField: (
unvalidatedData: unknown,
field: string,
) => Promise<ValidateFieldResult>;
};
In order to make an adapter for your validation library of choice, you can create a function that accepts a schema from the validation library and turns it into a validator.
Note the use of createValidator
.
It takes care of unflattening the data for nested objects and arrays
since the form doesn't know anything about object and arrays and this should be handled by the adapter.
For more on this you can check the implementations for withZod
and withYup
.
The out-of-the-box support for yup
in this library works like this:
export const withYup = <Schema extends AnyObjectSchema>(
validationSchema: Schema,
// For best result with Typescript,
// we should type the `Validator`
// we return based on the provided schema
): Validator<InferType<Schema>> =>
createValidator({
validate: async (unvalidatedData) => {
// Validate with yup and return the
// validated & typed data or the error
if (isValid)
return {
data: { field1: "someValue" },
error: undefined,
};
else
return {
error: { field1: "Some error!" },
data: undefined,
};
},
validateField: async (unvalidatedData, field) => {
// Validate the specific field with yup
if (isValid) return { error: undefined };
else return { error: "Some error" };
},
});