Automatically generate a validator using JSON Schema and AJV for any TypeScript type.
Define a type in src/Example.ts
, e.g.:
export default interface ExampleType {
value: string;
/**
* @TJS-format email
*/
email?: string;
/**
* @default 42
*/
answer: number;
}
To generate a validator, run:
npx @rkesters/typescript-json-validator src/Example.ts ExampleType
This will generate src/Example.validator.ts
, which you can use:
import { readFileSync } from 'fs';
import validate from './Example.validator.ts';
const value: unknown = JSON.parse(readFileSync(process.argv[2], 'utf8'));
// this will through a clear error if `value` is not of the
// correct type. It will also fill in any default values
const validatedValue = validate(value);
console.log(validatedValue.value);
Note that types will be validated automatically, but you can also use annotations to add extra runtime checks, such as e-mail formatting. For annotations see: https://github.com/YousefED/typescript-json-schema#annotations
Usage: typescript-json-schema <path-to-typescript-file> <type>
Options:
--help Show help [boolean]
--version Show version number [boolean]
--refs Create shared ref definitions.
[boolean] [default: true]
--topRef Create a top-level ref definition.
[boolean] [default: false]
--noExtraProps Disable additional properties in objects by default.
[boolean] [default: true]
--propOrder Create property order definitions.
[boolean] [default: true]
--strictNullChecks Make values non-nullable by default.
[boolean] [default: true]
--id ID of schema. [string] [default: ""]
--uniqueItems Validate `uniqueItems` keyword
[boolean] [default: true]
--unicode calculate correct length of strings with unicode pairs
(true by default). Pass false to use .length of
strings that is faster, but gives "incorrect" lengths
of strings with unicode pairs - each unicode pair is
counted as two characters. [boolean] [default: true]
--nullable support keyword "nullable" from Open API 3
specification. [boolean] [default: true]
--format formats validation mode ('fast' by default). Pass
'full' for more correct and slow validation or false
not to validate formats at all. E.g., 25:00:00 and
2015/14/33 will be invalid time and date in 'full'
mode but it will be valid in 'fast' mode.
[choices: "fast", "full"] [default: "fast"]
--coerceTypes Change data type of data to match type keyword. e.g.
parse numbers in strings [boolean] [default: false]
--collection Process the file as a collection of types, instead of
one single type. [boolean] [default: false]
-o, --out name of output, optional [string]
MIT