realtime + local turn detection: drop the user-transcript wait

Add the configuration surface to drive a realtime service like Gemini
Live from local turn detection without paying user-transcript latency.
Cascaded pipelines wait for a transcript before ending the user's turn
because the downstream LLM needs the user's words recorded in context
— but that wait is pure latency in pipelines using local turn
detection to drive a realtime service, which consumes user audio
directly.

Set `wait_for_transcript_to_end_user_turn=False` on
`LLMUserAggregatorParams` to turn this on. With that single flag the
aggregator:

- drops `TranscriptionUserTurnStartStrategy` from the start strategies
  (so late-arriving realtime transcripts don't trigger new turns),
- sets `wait_for_transcript=False` on any stop strategy that supports
  it (so the turn ends on the audible end of the turn, without
  waiting for a transcript),
- fires `on_user_turn_stopped` on the audible end of the turn with
  empty `content` (since the transcript hasn't arrived), and
- defers the context flush until the transcript arrives or a backstop
  timer fires.

A new `on_user_turn_message_finalized` event fires when the user's
message has been written to context. In the default mode it
coincides with `on_user_turn_stopped`; in the delayed-transcript mode
it fires later. Consumers that want the populated transcript should
subscribe to `on_user_turn_message_finalized` — it's the event that
always carries the user message, regardless of mode.

Strategy mutations are logged: loudly when the user passed their own
strategies (we're overwriting parts of their config), quietly
otherwise. The strategy-level `wait_for_transcript` parameter on
`TurnAnalyzerUserTurnStopStrategy` and `SpeechTimeoutUserTurnStopStrategy`
remains exposed for advanced cases.

The example `realtime-gemini-live-local-vad.py` demonstrates the full
pattern.

changelog/4480.added.md

@@ -0,0 +1 @@
+- Added the configuration surface to drive a realtime service like Gemini Live from local turn detection without paying user-transcript latency. Cascaded pipelines wait for a transcript before ending the user's turn because the downstream LLM needs the user's words recorded in context — but that wait is pure latency in pipelines using local turn detection to drive a realtime service, which consumes user audio directly. Set `wait_for_transcript_to_end_user_turn=False` on `LLMUserAggregatorParams` to turn this on: the aggregator drops `TranscriptionUserTurnStartStrategy` from the start strategies (so late-arriving realtime transcripts don't trigger new turns), sets `wait_for_transcript=False` on any stop strategies that support it (so the turn ends without waiting for a transcript), fires `on_user_turn_stopped` on the audible end of the turn (with empty content, since the transcript hasn't arrived), and defers the context flush until the transcript arrives. A new `on_user_turn_message_finalized` event fires when the user's message has been written to context — in the default mode it coincides with `on_user_turn_stopped`; in the delayed-transcript mode it fires later. Consumers that want the populated transcript should subscribe to `on_user_turn_message_finalized`. See `examples/realtime/realtime-gemini-live-local-vad.py` for the full pattern.

changelog/4480.fixed.md

@@ -1 +0,0 @@
-- `TurnAnalyzerUserTurnStopStrategy` and `SpeechTimeoutUserTurnStopStrategy` no longer always wait for a transcript before ending the user turn. The wait is still correct for cascaded pipelines, where the downstream LLM can't respond until the user's words are recorded in its context — but it's pure latency in pipelines using local turn detection to drive a realtime service like Gemini Live. The strategies now infer from the presence of an `STTMetadataFrame` whether a transcript is needed, and accept a new `require_transcript: bool | None` parameter to override the heuristic explicitly.