Add real-time speaker diarization support

- New whisper_live/diarization.py: SpeakerDiarizer with online clustering
- Uses pyannote.audio speaker embeddings (optional dependency)
- Cosine similarity threshold for speaker matching (default 0.55)
- Running average embedding update for speaker stability
- Configurable max_speakers limit (default 10)
- Client options: enable_diarization, max_speakers
- Segments include 'speaker' field when diarization is active
- Graceful fallback: logs warning if pyannote not installed
- Added 12 unit tests (mock-based, no GPU required)

whisper_live/backend/faster_whisper_backend.py

@@ -35,6 +35,7 @@ class ServeClientFasterWhisper(ServeClientBase):
         cache_path="~/.cache/whisper-live/",
         translation_queue=None,
         hotwords=None,
+        diarization=None,
     ):
         """
         Initialize a ServeClient instance.
@@ -64,7 +65,8 @@ class ServeClientFasterWhisper(ServeClientBase):
             no_speech_thresh,
             clip_audio,
             same_output_threshold,
-            translation_queue
+            translation_queue,
+            diarization,
         )
         self.cache_path = cache_path
         self.model_sizes = [