JAM is a utility layer for assisting in converting JSON API payloads into a convenient client side format, and in converting back to server compatible JSON API payloads.
A schema describing the server side model format (typically automatically generated) allows:
- Data type conversions (e.g. incoming timestamps are converted using
moment
). - Relationship type checking.
- Basic validation (e.g. required fields).
- Prefill with default values.
- Provision of select field options.
npm install js-jam
or
yarn add js-jam
If you happen to be using js-tinyapi
for your API client, a convenient middleware
is provided to allow automatic JSON API conversions:
import {jamMiddleware} from 'js-jam'
api = new Api()
api.addMiddleware(jamMiddleware)
api.listMovies() // returns flattened data instead of JSON API
See js-tinyapi
's documentation for more details.
Before data can be manipulated a schema describing the structure of the data must be defined. There are a number of ways to do it, the two most common are to define the data manually, or import it automatically using an external package.
Schemas are built using the Schema
class:
import {Schema} from 'js-jam'
let schema = new Schema()
To define models in a schema, use the merge
method, which accepts an object
argument describing a part of a schema:
schema.merge({})
merge
may be called any number of times. Each subsequent call will overwrite
any overlapping models.
The structure of the schema object is similar in some ways to the structure of a JSON-API object. Take for example the following definition of a movie:
{
movie: {
attributes: {
name: {
required: true
},
year: {}
},
relationships: {
actors: {
type: "person",
many: true
}
}
},
person: {
attributes: {
name: {
required: true
}
}
}
}
This defines two models: movie
and person
.
Options for attributes are currently limited to required
.
Options for relationships:
- type
- required
- many
If you're using Django and DRF, your schema can be loaded into JAM automatically, which is particularly convenient.
Refer to Django-JAM
Once data has been loaded from your server, conversion to a local format
is achieved via a call to Scema's method fromJsonApi
:
import schema from 'mySchema'
const jsonApiData = fetch('/movies/?include=actors')
/*
jsonApiData: {
data: [
{
type: 'movie',
id: 1,
attributes: {
title: 'Rocky'
},
relationships: {
actors: {
data: [
{
type: 'person',
id: 1
},
{
type: 'person',
id: 2
}
]
}
}
},
{
type: 'movie',
id: 2,
attributes: {
title: 'Rocky 2'
},
relationships: {
actors: {
data: [
{
type: 'person',
id: 1
}
]
}
}
}
}
]
included: [
{
type: 'person',
id: 1,
attributes: {
name: 'Sylvester Stalone'
}
},
{
type: 'person',
id: 2,
attributes: {
name: 'Dolf Lundrem'
}
}
]
*/
const data = schema.fromJsonApi(jsonApiData)
/*
data: [
{
_type: 'movie',
id: 1,
title: 'Rocky'
actors: [
{
_type: 'person',
id: 1,
name: 'Sylvester Stalone'
},
{
_type: 'person',
id: 2,
name: 'Dolf Lundrem'
}
]
},
{
_type: 'movie',
id: 2,
title: 'Rocky 2'
actors: [
{
_type: 'person',
id: 1,
name: 'Sylvester Stalone'
}
]
}
]
*/
schema.toJsonApi(data)
Note that when linking relationships together each instance of a resource is one and the same. So, in the above example, both instances of the person resource with ID 1 are actually the same JavaScript object.
Below we'll put together the above snippets into working examples. All examples use the following JSON schema file:
// schema.json
{
"movie": {
"attributes": {
"name": {
"required": true
},
"year": {}
},
"relationships": {
"actors": {
"type": "person",
"many": true
}
}
},
"person": {
"attributes": {
"name": {
"required": true
}
}
}
}
import Schema from 'js-jam'
const schema = new Schema(require('./schema.json'))
const data = schema.fromJsonApi(loadMyData())
import Schema, {jamMiddleware} from 'js-jam'
import Api from 'js-tinyapi'
const api = new Api(require('./api.json'))
api.addMiddleware(jamMiddleware)
const schema = new Schema(require('./schema.json'))
const data = api.moviesList()