fix: Open WebUI integration — Harmony stripping, VRAM eviction, concurrency lock

- Add harmony.py: strip GPT-OSS-20B analysis/thinking channel from both
  streaming and non-streaming responses (HarmonyStreamFilter + extract_final_text)
- Add per-model asyncio.Lock in llamacpp backend to prevent concurrent C++
  access that caused container segfaults (exit 139)
- Fix chat handler swap for streaming: move inside _stream_generate within
  lock scope (was broken by try/finally running before stream was consumed)
- Filter /v1/models to return only LLM models (hide ASR/TTS from chat dropdown)
- Correct Qwen3.5-4B estimated_vram_gb: 4 → 9 (actual allocation ~8GB)
- Add GPU memory verification after eviction with retry loop in vram_manager
- Add HF_TOKEN_PATH support in main.py for gated model access
- Add /v1/audio/models and /v1/audio/voices discovery endpoints (no auth)
- Add OOM error handling in both backends and chat route
- Add AUDIO_STT_SUPPORTED_CONTENT_TYPES for webm/wav/mp3/ogg
- Add performance test script (scripts/perf_test.py)
- Update tests to match current config (42 tests pass)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

kischdle/llmux/tests/test_model_registry.py

@@ -10,12 +10,12 @@ def registry():
 
 def test_list_virtual_models(registry):
     models = registry.list_virtual_models()
-    assert len(models) == 15
+    assert len(models) == 12  # only LLM models, not ASR/TTS
     names = [m["id"] for m in models]
     assert "Qwen3.5-9B-FP8-Thinking" in names
     assert "GPT-OSS-20B-High" in names
-    assert "cohere-transcribe" in names
-    assert "Chatterbox-Multilingual" in names
+    assert "cohere-transcribe" not in names
+    assert "Chatterbox-Multilingual" not in names
 
 
 def test_virtual_model_openai_format(registry):
@@ -28,7 +28,7 @@ def test_virtual_model_openai_format(registry):
 def test_resolve_virtual_to_physical(registry):
     physical_id, physical, params = registry.resolve("Qwen3.5-9B-FP8-Thinking")
     assert physical_id == "qwen3.5-9b-fp8"
-    assert physical.backend == "transformers"
+    assert physical.backend == "llamacpp"
     assert params == {"enable_thinking": True}
 
 
@@ -58,7 +58,7 @@ def test_resolve_unknown_model_raises(registry):
 def test_get_physical(registry):
     physical = registry.get_physical("qwen3.5-9b-fp8")
     assert physical.type == "llm"
-    assert physical.estimated_vram_gb == 9
+    assert physical.estimated_vram_gb == 10
 
 
 def test_get_physical_unknown_raises(registry):