Skip to content

capaj/moonridge-client

Repository files navigation

Moonridge-client

client library for Mongo remote ORM framework Moonridge.

Install with jspm(for browser) or npm(for node):

npm install moonridge-client

jspm install moonridge-client

Is JSPM really needed?

Moonridge is written in commonJS format(client works in node and browser without much hassle), so you need globally installed jspm to be able to install it with one command. For loading in the browser, you need systemJS(which JSPM installs for you). If you don't have jspm, install it with this command:

npm i jspm -g

As it works in node, it should be possible to run in browserify/webpack as well, but I don't guarantee it. JSPM might also be able to make you a bundle of all scripts needed. This bundle can then be used as regular script file with a script tag. Theoretically.

Basic usage

// ES6 import
import MR from 'moonridge-client'
/// cjs
const MR = require('moonridge-client')

// then just connect to your backend
const backend = MR({url: 'http://localhost:9000'})
// connecting with sync authorization
const backend = MR({url: 'http://localhost:9000', hs: { //hs stands for handshake(is passed to socket.io's connect method)
  query: 'email=john@mail.com&password=' + encodeURIComponent(password)
}})

// finally declare which model are you going to use(models are defined with schemas on the backend)
const bookModel = backend.model('book')

Backend instance API

model(string)

synchronously defines a model to be used with the backend

@param {String} name
@returns {Model}

fetchAllModels()

asynchronously gets a list of all models on the backend and then synchronously defines them

@param {String} name
@returns {Model}

authorize(object)

async authorization-use when user connects as anonymous and then gains some priviliges while session is opened

@param {Object} auth object
@returns {Promise} resolved on succesful authorization

deAuthorize()

user logged out/lost priviliges

@returns {Promise} resolved when user was removed from this socket on the server

Model instance API

query()

@returns {QueryChainable} which has same methods as mongoose.js query. When you chain all query
                          conditions, you use exec() to fire the query, Promise is returned

liveQuery([previousLiveQuery])

@param {Object} [previousLQ] useful when we want to modify an existing running LQ, pass it after it is stopped
@returns {QueryChainable} same as query, difference is that executing this QueryChainable won't return
                          promise, but liveQuery object itself

create(object)

@param {Object} toCreate
@returns {Promise} resolved when object is created in the database

update(object)

@param {Object} toUpdate moonridge object
@param {Boolean} resolveWhole if true, it will resolve with whole object instead of just the version
@returns {Promise} resolved when object is saved

Why not return whoole object by default? Because it would waste bandwith. If you need to display latest version of a document, it is preferable to use liveQuery().findOne({_id: id}), instead of manually updating your copy

remove(toRemove)

@param {Object|String} toRemove must have and _id if an object, otherwise we assume the string is the id
@returns {Promise}

addToSet(query, path, item)

@param {Object} query which will be used to find one document to update
@param {String} path
@param {*} item it is highly recommended to use simple values, not objects
@returns {Promise} resolved when object is updated

removeFromSet(query, path, item)

@param {Object} query which will be used to find one document to update
@param {String} path
@param {*} item it is highly recommended to use simple values, not objects
@returns {Promise} resolved when object is updated

on()

@returns {Promise} resolved when subscription is created on the server

off()

@returns {Promise} resolved when subscription is unsubscribed on the server

getSchema()

@returns {Promise} resolved with serialized and deserialized mongoose schema using mongoose-schema-serializer

QueryChainable instance API

Mimics mongoose query api, with some methods dropped. Implemented methods are:

all and box circle comment count distinct elemMatch equals exists find findOne geometry gt gte hint in intersects limit lt lte maxDistance maxScan mod ne near nin nor or ne polygon populate read regex select size skip slice sort where within

js-standard-style