from __future__ import annotations
import contextlib
import pytest
from fastmcp import Client
from mcp_agent_mail.app import build_mcp_server
class _StubOut:
def __init__(self, text: str):
self.content = text
self.model = "m"
self.provider = "p"
@pytest.mark.asyncio
async def test_summarize_threads_llm_refinement(isolated_env, monkeypatch):
# Force LLM enabled
from mcp_agent_mail import config as _config
monkeypatch.setenv("LLM_ENABLED", "true")
with contextlib.suppress(Exception):
_config.clear_settings_cache()
# Monkeypatch LLM call to return JSON content the app will parse
from mcp_agent_mail import app as app_mod
async def _fake_complete(*_a, **_k):
return _StubOut(
'{"threads": [{"thread_id": "T-1", "key_points": ["refined"], "actions": ["do"]}, {"thread_id": "T-2", "key_points": ["also refined"], "actions": ["act"]}], "aggregate": {"top_mentions": [], "key_points": ["K"], "action_items": ["A"]}}'
)
monkeypatch.setattr(app_mod, "complete_system_user", _fake_complete)
server = build_mcp_server()
async with Client(server) as client:
await client.call_tool("ensure_project", {"human_key": "/backend"})
await client.call_tool(
"register_agent",
{"project_key": "Backend", "program": "x", "model": "y", "name": "BlueLake"},
)
# Seed two threads to trigger multi-thread mode
for tid in ("T-1", "T-2"):
await client.call_tool(
"send_message",
{
"project_key": "Backend",
"sender_name": "BlueLake",
"to": ["BlueLake"],
"subject": f"Msg in {tid}",
"body_md": "body",
"thread_id": tid,
},
)
# Use comma-separated thread_id for multi-thread mode
res = await client.call_tool(
"summarize_thread",
{"project_key": "Backend", "thread_id": "T-1,T-2", "llm_mode": True, "per_thread_limit": 5},
)
payload = res.data
assert payload.get("threads")
# Ensure LLM-refined aggregate keys present
agg = payload.get("aggregate") or {}
assert agg.get("action_items") == ["A"]
