pipecat instrumentation refactor

[vipyne]

cc @duncankmckinnon

---

Note

Refactors the Pipecat OpenInference observer to use turn-tracking with expanded frame handling and precise span lifecycle/metrics, and updates LLM model name extraction to prefer _full_model_name.

• Observer refactor (_observer.py):
  • Change OpenInferenceObserver to inherit from TurnTrackingObserver and UserBotLatencyLogObserver; remove custom turn state and delegate to super.
  • Expand handled frames (LLM*, TTS*, VADUser*, MetricsFrame, etc.) and route through _handle_service_frame.
  • Start service spans on VADUserStartedSpeakingFrame (STT), LLMFullResponseStartFrame (LLM), and TTSStartedFrame (TTS); finish spans on EndFrame, CancelFrame, ErrorFrame, LLMFullResponseEndFrame, BotStoppedSpeakingFrame, and STT completion after VADUserStoppedSpeakingFrame + final TranscriptionFrame.
  • Accumulate user (TranscriptionFrame) and bot (TTSTextFrame, LLMTextFrame) text; set input.value/output.value on span/turn with spacing based on includes_inter_frame_spaces.
  • Process MetricsFrame to set attributes (e.g., service.processing_time_seconds) and compute span end times; attach conversation.user_to_bot_latency.
  • Create service spans as children of pipecat.conversation.turn; update turn start/end (_start_turn, _end_turn) and attributes (conversation.turn_number, duration, end reason).
• Attributes (_attributes.py):
  • Update LLMServiceAttributeExtractor to prefer service._full_model_name for llm.model_name, falling back to model_name/model.

^{Written by Cursor Bugbot for commit 0232542. This will update automatically on new commits. Configure here.}

[github-actions]

CLA Assistant Lite bot:
Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request}