added rust latency measurement example

[xianshijing-lk]

Setup (Note, replace with your own endpoint and token from "lk token create" cmd)
export LIVEKIT_URL=wss://xianstaging-hixkk74p.staging.livekit.cloud # Note, change to your own endpoint
export LIVEKIT_TOKEN=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE3NzYzMjcyMDQsImlkZW50aXR5Ijoic3hpYW4iLCJpc3MiOiJBUElHejJLUXU0UGJ6YkEiLCJuYW1lIjoic3hpYW4iLCJuYmYiOjE3NzAzMjcyMDQsInN1YiI6InN4aWFuIiwidmlkZW8iOnsicm9vbSI6ImNwcCIsInJvb21Kb2luIjp0cnVlfX0.oI0DCLyCMK-y6yxyQ3cFPArGJa03XHeeUrC_t2LUCgU

Rust the test with benchmark and logging:
RUST_LOG=info cargo run -p agent_audio_latency -- \
--url "$LIVEKIT_URL"
--token "$LIVEKIT_TOKEN"
--benchmark
--user-speech-threshold-dbfs=-50
--speaker-speech-threshold-dbfs=-55
--speaker-confirm-ms=120

Run the pure tests without speech latency detection (this will reduce the unnecessary test processing overhead):
RUST_LOG=info cargo run -p agent_audio_latency -- \
--url "$LIVEKIT_URL"
--token "$LIVEKIT_TOKEN"