website·Indexed 34 days ago
https://deepstream.io/docs/A real-time data synchronization and messaging framework. Includes guides for real-time search, WebRTC full-mesh networking, and comprehensive API documentation for Java v2 and JavaScript clients, covering records, lists, RPCs, and event handling.
deepstream is a platform that enables clients and backend services to synchronize data and exchange events with a strong emphasis on security. It allows for server-side configuration without requiring custom code to connect to caches, databases, and message buses. Key features include:
AnonymousRecord is a record wrapper without a predefined name. It allows a developer to swap the underlying record it is bound to without breaking existing bindings. This is particularly useful for populating user interfaces with data selected from a list of entries.@Injectable service. This allows multiple components to share the same connection instance and provides a centralized place for common operations like logging in or fetching records.import { Injectable } from '@angular/core';
@Injectable()
export class DsService {
private ds;
public dsInstance;
constructor() {
this.ds = this.dsInstance = new DeepstreamClient('localhost:6020');
}
login (credentials?, loginHandler?) {
this.ds.login(credentials, loginHandler);
}
getRecord(name) {
return this.ds.record.getRecord(name);
}
getList(name){
return this.ds.record.getList(name);
}
}
user object provides information about the user attempting an action. user.id contains the username provided at login, and user.data contains metadata (such as roles or flags) returned by the authentication webhook or file-based auth. The now variable can be used alongside user.data for time-based validation.record:
"*":
# Allow access only if 24 hours have passed since signup
record: "user.data.timestamp + 24*60*60*1000 < now"
write: "user.data.timestamp + 24*60*60*1000 < now"ws-json endpoint allows users to interact with Deepstream using JSON serialized text blobs instead of the default Protobuf layer. This is primarily intended for developers building SDKs who wish to avoid the complexity of Protobuf.The DeepstreamRuntimeErrorHandler interface allows developers to handle common runtime errors centrally, reducing the need for repetitive try/catch blocks. It is primarily triggered by asynchronous errors such as timeouts or merge conflicts.
Note: Request-specific errors (e.g., an RPC timeout or a record permission error) are not handled here; they are passed directly to the requesting method or caught by a more specific listener.
Deepstream provides an MQTT endpoint that supports MQTT authentication (username/password), record retention, and QoS 1 write acknowledgments. Because MQTT uses a single concept for messaging, Deepstream bridges both events and records together. Subscribing to a topic (e.g., 'weather/london') will deliver updates from both event.emit() and record.setData() calls.
Key MQTT mapping:
retain: true: Tells Deepstream the message is a record, which will be saved in the database.qos: 1: Tells Deepstream to provide a write acknowledgment confirmation.const mqtt = require('mqtt')
const client = mqtt.connect('mqtt://localhost:1883', {
username: 'my-device-name',
password: 'my-device-password',
properties: {
authenticationMethod: 'a-custom-authentication-method',
authenticationData: JSON.stringify({ property: "another value" })
}
})
client.on('connect', function () {
client.subscribe('start/weather/london', (err) => console.log('subscribed'))
client.subscribe('stop/weather/london', (err) => console.log('subscribed'))
})
client.on('message', function (topic, message) {
if (topic === 'start-publishing') {
setInterval(() => {
client.publish('weather/london',
JSON.stringify({ temperature: 12 }),
{
retain: true, // Saves as a record in the database
qos: 1 // Requests write ack
},
console.log
)
}, 1000)
}
})