ephery

Ephery

Ephery is a dead simple, in-memory, client-side, fake API service. This tool is useful for prototyping client-side applications without hijacking Ajax. In particular, this tool was built to interact with React + Redux applications that rely on an API service, but Ephery can be useful for any framework that isn't too opinionated on data-fetching.

Getting started

npm install --save ephery

Define some basic schemas and fixtures.

// src/services/store.js
import Store from 'ephery';

const schemas = {
  user: {
    attributes: {
      name: true,
      email: true,
    },
    relationships: {
      blogs: {
        hasMany: 'blog',
      },
      comments: {
        hasMany: 'comment',
      },
    },
  },
  blog: {
    attributes: {
      title: true,
      createdAt: true,
    },
    relationships: {
      author: {
        belongsTo: 'user',
      },
      comments: {
        hasMany: 'comment',
      },
    },
  },
  comment: {
    attributes: {
      text: true,
      createdAt: true,
    },
    relationships: {
      user: {
        belongsTo: 'user',
      },
      blog: {
        belongsTo: 'blog',
      },
    },
  },
};

const fixtures = {
  user: {
    1: {
      name: 'John Doe',
      email: '[e-mail]',
      blogs: ['1'],
      comments: ['1'],
    },
  },
  blog: {
    1: {
      title: 'Tabs vs. Spaces',
      author: '1',
      comments: ['1'],
    },
  },
  comment: {
    1: {
      text: 'First comment!',
      user: '1',
      blog: '1',
    },
  },
};

export default new Store(schemas, fixtures);

Then somewhere in your application, you can invoke Ephery. All returned data is deeply nested JSON.

import api from '../services/store';

api.fetch('user', '1').then(user => {
  /*
  {
    name: 'John Doe',
    email: '[e-mail]',
    blogs: [{
      id: '1',
      title: 'Tabs vs. Spaces',
      author: '1',
      comments: ['1'],
    }],
    comments: [{
      id: '1',
      text: 'First comment!',
      user: '1',
      blog: '1',
    }]
  }
   */
});

API

.create(type, record) -> Object async

api.create('user', {
  name: 'Dylan',
  email: '[e-mail]',
}).then(user => {

});

.fetch(type, id) -> Object async

api.fetch('user', '1').then(user => {

});

.find(type, filters = {}) -> Object[] async

api.find('user', {
  name: 'Dylan',
}).then(users => {

});

.update(type, id, data) -> Object async

api.update('user', '1', {
  name: 'Bob',
}).then(user => {

});

.del(type, id) -> Boolean async

api.del('user', '1').then(success => {
  // true or false
});

React + Redux Example

Dan Abramov has an excellent course on Egghead.io titled "Idiomatic Redux," where he mocks a simple API service that is invoked by actions. Ephery can act as that service in applications that have more intensive CRUD requirements. Because Ephery returns deeply-nested JSON objects, responses can be easily normalized and merged with the state tree.

You'll need to create your own production-grade API service once your backend is finished, but Ephery can act as a drop-in replacement until then.

Current Gotchas

Relationships do not cascade right now. Meaning, if you delete a user entity that a blog's author relationship (a belongsTo relationship) points to, the blog is not deleted. I'm going to add this functionality soon. Also, if you're using Normalizr, you'll essentially have to duplicate schema definitions.

Also, the coverage could be a lot better, so more tests are coming as well.