feat(rag): overlapping chunker, retrieval knobs, sqlite-vec index

- Chunker: char overlap across boundaries + hard-split of oversized paragraphs.
- Retrieval knobs: rag_top_k / rag_min_score in settings + Settings UI.
- Vector index: sqlite-vec ANN over document embeddings, dual-written and
  backfilled, with brute-force cosine as the guaranteed fallback.

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

synapse/main.py

@@ -286,7 +286,11 @@ async def chat_stream_endpoint(payload: Dict[str, Any]):
             system_prompt = (system_prompt + separator + memory_block) if system_prompt else memory_block
 
         # Retrieve relevant uploaded documents (RAG) and inject the top chunks.
-        doc_hits = await store.search_documents(message, get_ollama_manager().embed, limit=3)
+        doc_hits = await store.search_documents(
+            message, get_ollama_manager().embed,
+            limit=app_settings.get("rag_top_k", 3),
+            min_score=app_settings.get("rag_min_score", 0.6),
+        )
         doc_titles: list = []
         if doc_hits:
             doc_block = "\n\n".join(f"[{d['title']}]\n{d['text']}" for d in doc_hits)