JTS - JSON Transformation Schema

A lib and a spec for document transformation, with extended JSON schema.

The Why

• To help in writing readable, declarative tranformations.

Example

/* A simple transformation. */
const jts = require('json-transformation-schema');
const data = {
	a: 1,
}
const transformation = {
	properties: {
		a: {
			type: 'number',
			targetType: 'string',
		},
		b: {
			prop: 'a',
			transform: (val) => val + 1,
		},
	},
}

jts.transformer(transformation).transform(data);
/* Gives
{
	a: "1",
	b: 2,
}

For more examples, check the examples dir.

# Or run the following (with node v8+):
$ node ./examples/cheatsheet.js

Installation

$ npm install json-transformation-schema

Use Cases

• Simple and quick document transofromations.

• Configurable trasnsformation.

Keys

• JTS focuses on transformations, where as most JSON Schema implementations focus on validation.

• JTS is a superset of JSON Schema.

• Structure the transformation document to resemble the final document, for easier development and maintenance.

Notes

• The tools doesn't intend to be a full fledged JSON Schema Validator, at least in the beginning.

• The tools was written to support another project doctable.

Development

Setup

$ sh ./setup.sh

ToDo

• Fill any empty test cases.

• Think of the config, operators, thich would be used process values before output. These could be strings or composable functions.

• Measure and improve performance.

• Documention.

  • JS-Doc comments to allow for IDE tips.
  • Formal documentation with usecases and examples.

• Fix the coverage number at 95%.

• Offer input validation.

• Compile the schemas into functions.

• Add the support for one or more JSON Schema RFCs.

• Separate the spec from the implementation. The spec would be just the basics, without addiotional functionality like defaults, type-customization, custm transform functions etc.Extendability should be provided through the implementation. IE: A pure spec schema, should be portable.

• Test against the official JSON Schema Test Suite.

• Benchmark the validator with JSON Schema Benchmark.

DevNotes

• For an example on extending jts, refer jts-templates.

• The reason for throwing errors, instead of collecting them are many:

  • It reduces the learning curve for new comers.
  • It makes extending the core easy.
  • It makes the implementation easier.
  • The lib concentrates on transformation, rather than validation.
  • If needed a collection mechanism could be added later, when the lib becomes stable, or when needed.