perf: fix model reload thrash and cut time-to-first-token

Ollama reloaded the chat model mid-message because plan/summary/router
calls sent divergent runner options (no num_ctx/num_batch) and omitted
keep_alive — the main cause of perceived slowness.

- share one keep-alive httpx.AsyncClient for all Ollama calls
- unify runner options (runner_options) + keep_alive on every model
  call, embeddings included
- drop the blocking LLM routing fallback (pure lexical heuristic)
- run RAG recall + plan + code-model pick in parallel inside the SSE
  stream, after the start event
- RAG cosine scoring off the event loop; cache /api/tags 30s and
  nvidia-smi 5s; frontend polls 2s->5s, warm poll backoff, dedup
  config fetch

feat: working session menu in TopBar (switch/create/rename/delete)
feat: workspace file deletion (DELETE /api/files + UI trash buttons)
docs: recommended Ollama env vars (KEEP_ALIVE, MAX_LOADED_MODELS...)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/app/main.py

@@ -14,6 +14,7 @@ from fastapi.staticfiles import StaticFiles
 
 from . import agent_config, coder, db, rag
 from .config import settings
+from .ollama_client import ollama
 from .routes import benchmark, chat, config, files, git, models, sessions, shell, system
 
 
@@ -48,6 +49,8 @@ async def lifespan(_: FastAPI):
     # Préchargement du modèle par défaut, sans bloquer le démarrage.
     asyncio.create_task(_warm_default_model())
     yield
+    # Ferme le pool HTTP partagé vers Ollama.
+    await ollama.aclose()
 
 
 app = FastAPI(