safe-data

safe-data is a simple schema-based library for data parsing and validating.

Installation

npm install --save safe-data

Basic usage

import { schema, rules } from 'safe-data';

// Declare a reminder schema.
const reminderSchema = schema.shape({
  isCompleted: schema.boolean(),
  title: schema.string({
    rules: [
      rules.minLength({ limit: 5 }),
      rules.maxLength({ limit: 64 }),
    ],
  }),
  notes: schema.string().nullish(),
  date: schema.string({
    rules: [rules.isoDatetime()],
  }).transform((value) => new Date(value)),
  priority: schema.enum([
    'low',
    'medium',
    'high',
  ]).optional(),
  tags: schema.array({
    item: schema.string(),
  }),
});

// Use the reminder schema to validate data.
// E.g. check HTML form data...
try {
  const form = document.getElementById('reminder-form');
  const formData = new FormData(form);
  const reminder = reminderSchema.validate({
    title: formData.get('title'),
    notes: formData.get('notes'),
    date: formData.get('date'),
    priority: formData.get('priority'),
    tags: formData.getAll('tags'),
  });
} catch (error) {
  console.error(error);
}
// ...or a server response.
fetchReminder()
  .then(reminderSchema.validate)
  .catch((error) => console.error(error));

Core concepts

The safe-data is divided into 2 main logical parts - schemas and rules. The schemas provide a way to validate core types - string, number, object, etc. While rules using for format validation, for example, to check that string is a valid URL.

• Schemas are created with schema object. Schema instances are immutable.
• Rules are just pure functions that accept a value validated by schema and throw an error if the value has an invalid format. You can use built-in safe-data rules or write yours own.

Optional and null values

All schemas are required by default. However, you can make use of optional, nullable, and nullish modifiers to mark schema as optional.

Null values

You can allow null values with nullable() method. This method wraps the original schema to NullableModifier. To get the wrapped schema use unwrap() method.

import { schema } from 'safe-data';

const stringSchema = schema.string();
const nullableString = stringSchema.nullable();

nullableString.validate(null); // returns null
nullableString.unwrap() === stringSchema; // true

Optional values

You can allow undefined values with optional() method. This method wraps the original schema to OptionalModifier. To get the wrapped schema use unwrap() method.

import { schema } from 'safe-data';

const stringSchema = schema.string();
const optionalString = stringSchema.optional();

optionalString.validate(undefined); // returns undefined
optionalString.unwrap() === stringSchema; // true

Null and optional values

You can allow null and undefined values with nullish() method. This method wraps the original schema to NullishModifier. To get the wrapped schema use unwrap() method.

import { schema } from 'safe-data';

const stringSchema = schema.string();
const nullishString = stringSchema.nullable();

nullishString.validate(null); // returns null
nullishString.validate(undefined); // returns undefined
nullishString.unwrap() === stringSchema; // true

Transform value after validation

You can use transform() method to transform value after validation. This method accepts transform function and wraps the original schema to ResultTransformer. To get the source schema use sourceSchema property.

import { schema, rules } from 'safe-data';

const isoDateSchema = schema.string({
  rules: [
    rules.isoDate(),
  ],
});
const eventDateSchema = isoDateSchema.transform((value) => new Date(value));

isoDateSchema.validate('2020-08-10'); // returns string
eventDateSchema.validate('2020-08-10'); // returns Date
eventDateSchema.sourceSchema === eventDateSchema;

TypeScript Support

safe-data supports TypeScript and ships with types in the library itself.

Also, you can use InferType to extract the validation result type.

import { schema, type InferType } from 'safe-data';

const reminderSchema = schema.shape({
  isCompleted: schema.boolean(),
  title: schema.string(),
  notes: schema.string().nullish(),
  priority: schema.enum(['low', 'medium', 'high'] as const),
});

type Reminder = InferType<typeof reminderSchema>;
// {
//   isCompleted: boolean;
//   title: string;
//   notes?: string | null | undefined;
//   priority: 'low' | 'medium' | 'high';
// }

Sometimes you need to get type before any transformations. For this use an InferInput type.

import { schema, rules, type InferInput } from 'safe-data';

const isoStringSchema = schema.string({
  rules: [
    rules.isoDate(),
  ],
});
const dateSchema = isoStringSchema.transform((value) => new Date(value));

type ISOString = InferInput<typeof dateSchema>; // string