Infrastructure components to traverse, build models and cause side effects based on a JSON Schema.
This can be leveraged to build a Declarative Driven Development framework where you build most of your app from a single source of truth (enriched JSON schemas) and keep it in sync on every update.
See Building Blocks for more details
Please check the tests and markdown *.md documentation files for in-depth API documentation (Note: needs update).
npm install json-schema-model-builder -Syarn add json-schema-model-builderYou can use it something like this...
import { schema } from "./schemas/person.json";
import { config } from "./config";
import {
createBuilder,
createState,
createDispatcher
} from "json-schema-model-builder";
const state = createState({ config });
const dispatcher = createDispatcher({ state, config });
const builder = createBuilder({ state, schema, dispatcher, config });
const built = builder.build();
const rendered = built.render();
// Render the darn thing!
console.log({ rendered });Note: The Builder and Renderer still need some work, but you can leverage and extend their current implementation already.
This project is a result of multiple efforts trying to generate source code or derive models from JSON schemas. There is a need for a framework with a set of well tested, battle-tested building blocks to make this much easier. The goal is to achieve true declarative model driven development.
Your domain and action models across your stack should automatically be derived from the same underlying declarative model (project blueprint).
In the near future we will likely add interactive graph driven development to the mix, using an approach like d3-force-graph-editor
You could in the end simply draw your full application as a directed graph and then assign sufficient meta data to nodes and edges. Then as you draw, have an asortment of pluggable generators generate most (if not all!) of the app for you!
As you reconfigure your graph your app would be auto-coded/generated to reflect the graph in real time!!
Ideally you would then sync the state of the graph with other developers in real time, either using something like AppSync or GraphQL backend or perhaps something like automerge-graph or Immer
Infrastructure delivered by library
Builder builds a set of entities identified by resolvers and dispatches events as neededDispatcher Dispatches select events to state for update of modelState builds collections of entities and a directed graph from incoming model eventsYour concerns for output target
Model: target model to be built from state can either subscribe to the State or process the State when builder has finishedRenderer Rendering can be done when target model has been builtThe following examples act to demonstrate how you might go about using this infrastructure to generate source code or models.
The types collected, such as type, enum etc are collected in the store entry of the config object passed in.
const config = {};
const result = createSchema({ schema, config });
const { store } = config;
console.log(store);Note: As soon as we have a solid foundation and working examples we will move these into their own projects and publish them on npm
Uses jest for unit testing.
Please help add more/better test coverage :)
$ npm run test:watch$ npm run test2018 Kristian Mandrup (CTO@Tecla5)
MIT