Files
rondevu-client/README.md

459 lines
12 KiB
Markdown

# Rondevu Client
[![npm version](https://img.shields.io/npm/v/@xtr-dev/rondevu-client)](https://www.npmjs.com/package/@xtr-dev/rondevu-client)
🌐 **Simple, high-level WebRTC peer-to-peer connections**
TypeScript/JavaScript client for Rondevu, providing easy-to-use WebRTC connections with automatic signaling, username-based discovery, and built-in reconnection support.
**Related repositories:**
- [@xtr-dev/rondevu-client](https://github.com/xtr-dev/rondevu-client) - TypeScript client library ([npm](https://www.npmjs.com/package/@xtr-dev/rondevu-client))
- [@xtr-dev/rondevu-server](https://github.com/xtr-dev/rondevu-server) - HTTP signaling server ([npm](https://www.npmjs.com/package/@xtr-dev/rondevu-server), [live](https://api.ronde.vu))
- [@xtr-dev/rondevu-demo](https://github.com/xtr-dev/rondevu-demo) - Interactive demo ([live](https://ronde.vu))
---
## Features
- **High-Level Wrappers**: ServiceHost and ServiceClient eliminate WebRTC boilerplate
- **Username-Based Discovery**: Connect to peers by username, not complex offer/answer exchange
- **Semver-Compatible Matching**: Requesting chat@1.0.0 matches any compatible 1.x.x version
- **Privacy-First Design**: Services are hidden by default - no enumeration possible
- **Automatic Reconnection**: Built-in retry logic with exponential backoff
- **Message Queuing**: Messages sent while disconnected are queued and flushed on reconnect
- **Cryptographic Username Claiming**: Secure ownership with Ed25519 signatures
- **Service Publishing**: Package-style naming (chat.app@1.0.0) with multiple simultaneous offers
- **TypeScript**: Full type safety and autocomplete
- **Configurable Polling**: Exponential backoff with jitter to reduce server load
## Install
```bash
npm install @xtr-dev/rondevu-client
```
## Quick Start
### Hosting a Service (Alice)
```typescript
import { RondevuService, ServiceHost } from '@xtr-dev/rondevu-client'
// Step 1: Create and initialize service
const service = new RondevuService({
apiUrl: 'https://api.ronde.vu',
username: 'alice'
})
await service.initialize() // Generates keypair
await service.claimUsername() // Claims username with signature
// Step 2: Create ServiceHost
const host = new ServiceHost({
service: 'chat.app@1.0.0',
rondevuService: service,
maxPeers: 5, // Accept up to 5 connections
ttl: 300000 // 5 minutes
})
// Step 3: Listen for incoming connections
host.events.on('connection', (connection) => {
console.log('✅ New connection!')
connection.events.on('message', (msg) => {
console.log('📨 Received:', msg)
connection.sendMessage('Hello from Alice!')
})
connection.events.on('state-change', (state) => {
console.log('Connection state:', state)
})
})
host.events.on('error', (error) => {
console.error('Host error:', error)
})
// Step 4: Start hosting
await host.start()
console.log('Service is now live! Others can connect to @alice')
// Later: stop hosting
host.dispose()
```
### Connecting to a Service (Bob)
```typescript
import { RondevuService, ServiceClient } from '@xtr-dev/rondevu-client'
// Step 1: Create and initialize service
const service = new RondevuService({
apiUrl: 'https://api.ronde.vu',
username: 'bob'
})
await service.initialize()
await service.claimUsername()
// Step 2: Create ServiceClient
const client = new ServiceClient({
username: 'alice', // Connect to Alice
serviceFqn: 'chat.app@1.0.0',
rondevuService: service,
autoReconnect: true,
maxReconnectAttempts: 5
})
// Step 3: Listen for connection events
client.events.on('connected', (connection) => {
console.log('✅ Connected to Alice!')
connection.events.on('message', (msg) => {
console.log('📨 Received:', msg)
})
// Send a message
connection.sendMessage('Hello from Bob!')
})
client.events.on('disconnected', () => {
console.log('🔌 Disconnected')
})
client.events.on('reconnecting', ({ attempt, maxAttempts }) => {
console.log(`🔄 Reconnecting (${attempt}/${maxAttempts})...`)
})
client.events.on('error', (error) => {
console.error('❌ Error:', error)
})
// Step 4: Connect
await client.connect()
// Later: disconnect
client.dispose()
```
## Core Concepts
### RondevuService
Handles authentication and username management:
- Generates Ed25519 keypair for signing
- Claims usernames with cryptographic proof
- Provides API client for signaling server
### ServiceHost
High-level wrapper for hosting a WebRTC service:
- Automatically creates and publishes offers
- Handles incoming connections
- Manages ICE candidate exchange
- Supports multiple simultaneous peers
### ServiceClient
High-level wrapper for connecting to services:
- Discovers services by username
- Handles offer/answer exchange automatically
- Built-in auto-reconnection with exponential backoff
- Event-driven API
### RTCDurableConnection
Low-level connection wrapper (used internally):
- Manages WebRTC PeerConnection lifecycle
- Handles ICE candidate polling
- Provides message queue for reliability
- State management and events
## API Reference
### RondevuService
```typescript
const service = new RondevuService({
apiUrl: string, // Signaling server URL
username: string, // Your username
keypair?: Keypair // Optional: reuse existing keypair
})
// Initialize service (generates keypair if not provided)
await service.initialize(): Promise<void>
// Claim username with cryptographic signature
await service.claimUsername(): Promise<void>
// Check if username is claimed
service.isUsernameClaimed(): boolean
// Get current username
service.getUsername(): string
// Get keypair
service.getKeypair(): Keypair
// Get API client
service.getAPI(): RondevuAPI
```
### ServiceHost
```typescript
const host = new ServiceHost({
service: string, // Service FQN (e.g., 'chat.app@1.0.0')
rondevuService: RondevuService,
maxPeers?: number, // Default: 5
ttl?: number, // Default: 300000 (5 minutes)
isPublic?: boolean, // Default: true
rtcConfiguration?: RTCConfiguration
})
// Start hosting
await host.start(): Promise<void>
// Stop hosting and cleanup
host.dispose(): void
// Get all active connections
host.getConnections(): RTCDurableConnection[]
// Events
host.events.on('connection', (conn: RTCDurableConnection) => {})
host.events.on('error', (error: Error) => {})
```
### ServiceClient
```typescript
const client = new ServiceClient({
username: string, // Host username to connect to
serviceFqn: string, // Service FQN (e.g., 'chat.app@1.0.0')
rondevuService: RondevuService,
autoReconnect?: boolean, // Default: true
maxReconnectAttempts?: number, // Default: 5
rtcConfiguration?: RTCConfiguration
})
// Connect to service
await client.connect(): Promise<RTCDurableConnection>
// Disconnect and cleanup
client.dispose(): void
// Get current connection
client.getConnection(): RTCDurableConnection | null
// Events
client.events.on('connected', (conn: RTCDurableConnection) => {})
client.events.on('disconnected', () => {})
client.events.on('reconnecting', (info: { attempt: number, maxAttempts: number }) => {})
client.events.on('error', (error: Error) => {})
```
### RTCDurableConnection
```typescript
// Connection state
connection.state: 'connected' | 'connecting' | 'disconnected'
// Send message (returns true if sent, false if queued)
await connection.sendMessage(message: string): Promise<boolean>
// Queue message for sending when connected
await connection.queueMessage(message: string, options?: QueueMessageOptions): Promise<void>
// Disconnect
connection.disconnect(): void
// Events
connection.events.on('message', (msg: string) => {})
connection.events.on('state-change', (state: ConnectionStates) => {})
```
## Configuration
### Polling Configuration
The signaling uses configurable polling with exponential backoff:
```typescript
// Default polling config
{
initialInterval: 500, // Start at 500ms
maxInterval: 5000, // Max 5 seconds
backoffMultiplier: 1.5, // Increase by 1.5x each time
maxRetries: 50, // Max 50 attempts
jitter: true // Add random 0-100ms to prevent thundering herd
}
```
This is handled automatically - no configuration needed.
### WebRTC Configuration
Provide custom STUN/TURN servers:
```typescript
const host = new ServiceHost({
service: 'chat.app@1.0.0',
rondevuService: service,
rtcConfiguration: {
iceServers: [
{ urls: 'stun:stun.l.google.com:19302' },
{
urls: 'turn:turn.example.com:3478',
username: 'user',
credential: 'pass'
}
]
}
})
```
## Username Rules
- **Format**: Lowercase alphanumeric + dash (`a-z`, `0-9`, `-`)
- **Length**: 3-32 characters
- **Pattern**: `^[a-z0-9][a-z0-9-]*[a-z0-9]$`
- **Validity**: 365 days from claim/last use
- **Ownership**: Secured by Ed25519 public key signature
## Examples
### Chat Application
See [demo/demo.js](./demo/demo.js) for a complete working example.
### Persistent Keypair
```typescript
// Save keypair to localStorage
const service = new RondevuService({
apiUrl: 'https://api.ronde.vu',
username: 'alice'
})
await service.initialize()
await service.claimUsername()
// Save for later
localStorage.setItem('rondevu-keypair', JSON.stringify(service.getKeypair()))
localStorage.setItem('rondevu-username', service.getUsername())
// Load on next session
const savedKeypair = JSON.parse(localStorage.getItem('rondevu-keypair'))
const savedUsername = localStorage.getItem('rondevu-username')
const service2 = new RondevuService({
apiUrl: 'https://api.ronde.vu',
username: savedUsername,
keypair: savedKeypair
})
await service2.initialize() // Reuses keypair
```
### Message Queue Example
```typescript
// Messages are automatically queued if not connected yet
client.events.on('connected', (connection) => {
// Send immediately
connection.sendMessage('Hello!')
})
// Or queue for later
await client.connect()
const conn = client.getConnection()
await conn.queueMessage('This will be sent when connected', {
expiresAt: Date.now() + 60000 // Expire after 1 minute
})
```
## Migration from v0.9.x
v0.11.0+ introduces high-level wrappers, RESTful API changes, and semver-compatible discovery:
**API Changes:**
- Server endpoints restructured (`/usernames/*``/users/*`)
- Added `ServiceHost` and `ServiceClient` wrappers
- Message queue fully implemented
- Configurable polling with exponential backoff
- Removed deprecated `cleanup()` methods (use `dispose()`)
- **v0.11.0+**: Services use `offers` array instead of single `sdp`
- **v0.11.0+**: Semver-compatible service discovery (chat@1.0.0 matches 1.x.x)
- **v0.11.0+**: All services are hidden - no listing endpoint
- **v0.11.0+**: Services support multiple simultaneous offers for connection pooling
**Migration Guide:**
```typescript
// Before (v0.9.x) - Manual WebRTC setup
const signaler = new RondevuSignaler(service, 'chat@1.0.0')
const context = new WebRTCContext()
const pc = context.createPeerConnection()
// ... 50+ lines of boilerplate
// After (v0.11.0) - ServiceHost wrapper
const host = new ServiceHost({
service: 'chat@1.0.0',
rondevuService: service
})
await host.start()
// Done!
```
## Platform Support
### Modern Browsers
Works out of the box - no additional setup needed.
### Node.js 18+
Native fetch is available, but WebRTC requires polyfills:
```bash
npm install wrtc
```
```typescript
import { WebRTCContext } from '@xtr-dev/rondevu-client'
import { RTCPeerConnection, RTCSessionDescription, RTCIceCandidate } from 'wrtc'
// Configure WebRTC context
const context = new WebRTCContext({
RTCPeerConnection,
RTCSessionDescription,
RTCIceCandidate
} as any)
```
## TypeScript
All types are exported:
```typescript
import type {
RondevuServiceOptions,
ServiceHostOptions,
ServiceHostEvents,
ServiceClientOptions,
ServiceClientEvents,
ConnectionInterface,
ConnectionEvents,
ConnectionStates,
Message,
QueueMessageOptions,
Signaler,
PollingConfig,
Credentials,
Keypair
} from '@xtr-dev/rondevu-client'
```
## License
MIT