feat: persona chat loop, web UI, and local (Ollama) embeddings

Phase 1 — persona + persistent memory chat loop:
- lyra/persona.py + personas/lyra.md: editable identity/voice (friend-first,
  honest, never invents poker math)
- lyra/chat.py: turn loop assembling persona + cross-session recall + recent
  context, persisting both sides to SQLite
- lyra/session.py, lyra/__main__.py: session lifecycle + `lyra` REPL

Phase 1.25 — reuse the old web UI:
- vendored the prior single-page UI into lyra/web/static, repointed to
  same-origin
- lyra/web/server.py (FastAPI): serves the UI and backs its endpoint contract
  (/v1/chat/completions, session CRUD, health, inert thinking-stream) with the
  new chat loop + memory; SQLite stays the single source of truth
- `lyra-web` console script

Local backends — test for free, no OpenAI key:
- llm.embed routes via EMBED_BACKEND (cloud=OpenAI, local=Ollama /api/embed)
- simplified UI backend selector to Local (Ollama) / Cloud (OpenAI), default local
- memory connection opened check_same_thread=False for the threaded server

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lyra/chat.py

@@ -0,0 +1,54 @@
+"""The chat turn loop: persona + recalled memory + recent context -> reply.
+
+Each turn assembles the persona system prompt, semantically-relevant memories
+recalled from across all past sessions, and the recent turns of the current
+session, then asks the model for a reply and persists both sides.
+"""
+from __future__ import annotations
+
+from lyra import llm, memory, persona
+from lyra.llm import Backend, Message
+
+RECALL_K = 5
+RECENT_N = 10
+
+
+def _memory_note(exchanges: list[memory.Exchange]) -> Message:
+    """Format recalled memories as a system note Lyra can draw on."""
+    lines = []
+    for ex in exchanges:
+        when = ex.created_at[:10]  # YYYY-MM-DD
+        lines.append(f"- ({when}, {ex.role}) {ex.content}")
+    body = "Relevant things you remember from past conversations:\n" + "\n".join(lines)
+    return {"role": "system", "content": body}
+
+
+def build_messages(session_id: str, user_msg: str) -> list[Message]:
+    """Assemble the full message list for one turn."""
+    messages: list[Message] = [{"role": "system", "content": persona.system_prompt()}]
+
+    recent = memory.recent(session_id, n=RECENT_N)
+    recent_ids = {ex.id for ex in recent}
+
+    # Cross-session recall, minus anything already shown in the recent window.
+    recalled = [
+        ex for ex in memory.recall(user_msg, k=RECALL_K) if ex.id not in recent_ids
+    ]
+    if recalled:
+        messages.append(_memory_note(recalled))
+
+    for ex in recent:
+        messages.append({"role": ex.role, "content": ex.content})
+
+    messages.append({"role": "user", "content": user_msg})
+    return messages
+
+
+def respond(session_id: str, user_msg: str, backend: Backend = "cloud") -> str:
+    """Produce Lyra's reply to a single user message and persist the exchange."""
+    messages = build_messages(session_id, user_msg)
+    reply = llm.complete(messages, backend=backend)
+
+    memory.remember(session_id, "user", user_msg)
+    memory.remember(session_id, "assistant", reply)
+    return reply