Implement RPC request batching and throttling

Adds automatic request batching to reduce HTTP overhead by combining
multiple RPC calls into a single request.

Features:
- RpcBatcher class for intelligent request batching
- Configurable batch size (default: 10 requests)
- Configurable wait time (default: 50ms)
- Throttling to prevent overwhelming the server (default: 10ms)
- Automatic flushing when batch is full
- Enabled by default, can be disabled via options

Changes:
- Created rpc-batcher.ts with RpcBatcher class
- Updated RondevuAPI to use batcher by default
- Added batching option to RondevuOptions
- Updated README with batching documentation
- Bumped version to 0.16.0

Example usage:
  // Default (batching enabled with defaults)
  const rondevu = new Rondevu({ apiUrl: 'https://api.ronde.vu' })

  // Custom batching settings
  const rondevu = new Rondevu({
    apiUrl: 'https://api.ronde.vu',
    batching: { maxBatchSize: 20, maxWaitTime: 100 }
  })

  // Disable batching
  const rondevu = new Rondevu({
    apiUrl: 'https://api.ronde.vu',
    batching: false
  })

This can reduce HTTP requests by up to 90% during intensive operations
like ICE candidate exchange.

🤖 Generated with Claude Code
https://claude.com/claude-code

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@xtr-dev/rondevu-client",
-  "version": "0.15.0",
+  "version": "0.16.0",
   "description": "TypeScript client for Rondevu with durable WebRTC connections, automatic reconnection, and message queuing",
   "type": "module",
   "main": "dist/index.js",