"""OpenAI Agents SDK + khora example - session memory via ``KhoraSession``.
Runs without Postgres, Neo4j, or an API key. The mock LLM patches
``litellm.acompletion`` / ``litellm.aembedding`` so the example is
hermetic. The khora fixture spins up an in-memory ``sqlite_lance``
backend in a tmp dir.
The example does NOT spin up a real ``agents.Runner`` - that would
require a live LLM. Instead it exercises the three khora primitives the
adapter exposes directly: ``KhoraSession`` (SessionABC contract),
``khora_recall_tool`` (FunctionTool factory), and ``KhoraMemoryHooks``
(RunHooks-shaped). Each is what an ``Agent`` would call into.
Kept deliberately small (single ``add_items`` write) so it finishes
well under the 30s CI smoke budget - every khora write triggers a full
extraction pipeline that retries 3x on the mock LLM's non-JSON output.
"""
from __future__ import annotations
import asyncio
import sys
from pathlib import Path
# Add repo root to sys.path so ``examples._helpers`` is importable when
# this script is run from its own directory (CI smoke loop does that).
_REPO_ROOT = Path(__file__).resolve().parents[3]
if str(_REPO_ROOT) not in sys.path:
sys.path.insert(0, str(_REPO_ROOT))
from examples._helpers import embedded_khora, install_mock_llm # noqa: E402
from khora.integrations.openai_agents import ( # noqa: E402
KhoraMemoryHooks,
KhoraSession,
khora_recall_tool,
)
async def main() -> None:
install_mock_llm()
async with embedded_khora() as kb:
namespace = await kb.create_namespace()
ns_id = namespace.namespace_id
# 1) Session - one turn is enough to demonstrate the SessionABC
# contract. Every khora write runs full extraction, so we keep
# the example light to fit the 30s CI smoke budget.
session = KhoraSession(kb=kb, namespace=ns_id, session_id="example-conv-1")
await session.add_items([{"role": "user", "content": "We picked PostgreSQL for the user DB."}])
items = await session.get_items()
print(f"Session has {len(items)} item(s); latest: {items[-1]['content']!r}")
# 2) Recall tool - closes over (kb, namespace, top_k). Construction
# is pure Python; no LLM I/O. An Agent would invoke it later.
tool = khora_recall_tool(kb=kb, namespace=ns_id, top_k=3)
print(f"Built recall tool: name={tool.name!r}")
# 3) Memory hooks - construct only. ``on_tool_end`` would normally
# fire from inside ``Runner.run(...)`` and persist the tool
# output. We skip the live call here to keep the example fast.
hooks = KhoraMemoryHooks(kb=kb, namespace=ns_id, app_id="example")
print(f"Built memory hooks: app_id={hooks.app_id!r}")
if __name__ == "__main__":
asyncio.run(main())
