Files
hermes-agent/tests/gateway/test_steer_command.py
kshitijk4poor ab2f4602de refactor: MessageEvent to gateway/platforms/event.py; ElicitationHandler takes a call_context thunk
Breaks the two import cycles that forced Protocol stand-ins in the F821 sweep, so the two
sites now name the real types.

gateway/platforms/event.py (new leaf): MessageType, ProcessingOutcome, MessageEvent moved
out of base.py verbatim. Their only dependency is gateway.session.SessionSource; base.py
imported helpers.py at module level, so helpers could not name MessageEvent. Now
TextBatchAggregator is typed by the real MessageEvent. 249 importers repointed
(`from gateway.platforms.base import` -> `.event`, preserving each import's layout);
gateway.platforms.__init__ re-exports from .event. The three revert-scheduled PLUGIN-COMPAT
pointers that named these symbols (gateway.slash_commands → MessageType, dingtalk → MessageType,
photon → ProcessingOutcome) and their COMPAT_MANIFEST rows now target gateway.platforms.event.
Docs updated: ADDING_A_PLATFORM.md, adding-platform-adapters.md (en + zh-Hans).

tools/mcp_tool_sampling.py: ElicitationHandler no longer holds a back-reference to its
MCPServerTask (mcp_tool imports sampling, so the task type cannot be named there). It only
ever read owner._pending_call_context, so it takes `call_context: Callable[[], Context | None]`
and MCPServerTask passes `lambda: self._pending_call_context`. The consent call is one
`functools.partial`, run directly or inside the captured Context.

ty on the 11 touched production files vs origin/main: 0 new diagnostics, 14 resolved.
(The one `source: SessionSource = None` diagnostic moves with the class; typing it Optional
exposes ~60 unguarded call sites — separate follow-up.)

Tests: tests/gateway + tests/plugins + tests/tools + touched files, 18,235 passed; the 31
failures reproduce identically on origin/main (macOS /private/tmp, systemd socket,
long-path fixtures, live-service tests).
2026-09-07 22:47:33 +05:30

180 lines
6.5 KiB
Python

"""Tests for the gateway /steer command handler.
/steer injects a user message into the agent's next tool result without
interrupting. The gateway runner must:
1. When an agent IS running → call ``agent.steer(text)``, do NOT set
``_interrupt_requested``, do NOT touch ``_pending_messages``.
2. When the agent is the PENDING sentinel → fall back to /queue
semantics (store in ``adapter._pending_messages``).
3. When no agent is active → strip the slash prefix and let the normal
prompt pipeline handle it as a regular user message.
"""
from __future__ import annotations
from datetime import datetime
import time
from types import SimpleNamespace
from unittest.mock import AsyncMock, MagicMock
import pytest
from gateway.config import GatewayConfig, Platform, PlatformConfig
from gateway.platforms.event import MessageEvent
from gateway.session import SessionEntry, SessionSource, build_session_key
def _make_source() -> SessionSource:
return SessionSource(
platform=Platform.TELEGRAM,
user_id="u1",
chat_id="c1",
user_name="tester",
chat_type="dm",
)
def _make_event(text: str, channel_context: str | None = None) -> MessageEvent:
return MessageEvent(
text=text,
source=_make_source(),
message_id="m1",
channel_context=channel_context,
)
def _make_runner(session_entry: SessionEntry):
from gateway.run import GatewayRunner
runner = object.__new__(GatewayRunner)
runner.config = GatewayConfig(
platforms={Platform.TELEGRAM: PlatformConfig(enabled=True, token="***")}
)
adapter = MagicMock()
adapter.send = AsyncMock()
adapter._pending_messages = {}
runner.adapters = {Platform.TELEGRAM: adapter}
runner._voice_mode = {}
runner.hooks = SimpleNamespace(emit=AsyncMock(), loaded_hooks=False)
runner.session_store = MagicMock()
runner.session_store.get_or_create_session.return_value = session_entry
runner.session_store.load_transcript.return_value = []
runner.session_store.has_any_sessions.return_value = True
runner._running_agents = {}
runner._running_agents_ts = {}
runner._pending_messages = {}
runner._pending_approvals = {}
runner._session_db = MagicMock()
runner._session_db.get_session_title.return_value = None
runner._reasoning_config = None
runner._provider_routing = {}
runner._fallback_model = None
runner._show_reasoning = False
runner._is_user_authorized = lambda _source: True
runner._set_session_env = lambda _context: None
runner._should_send_voice_reply = lambda *_args, **_kwargs: False
runner._send_voice_reply = AsyncMock()
runner._capture_gateway_honcho_if_configured = lambda *args, **kwargs: None
runner._emit_gateway_run_progress = AsyncMock()
return runner, adapter
def _session_entry() -> SessionEntry:
return SessionEntry(
session_key=build_session_key(_make_source()),
session_id="sess-1",
created_at=datetime.now(),
updated_at=datetime.now(),
platform=Platform.TELEGRAM,
chat_type="dm",
total_tokens=0,
)
@pytest.mark.asyncio
async def test_steer_calls_agent_steer_and_does_not_interrupt():
"""When an agent is running, /steer must call agent.steer(text) and
leave interrupt state untouched."""
runner, adapter = _make_runner(_session_entry())
sk = build_session_key(_make_source())
running_agent = MagicMock()
running_agent.steer.return_value = True
runner._running_agents[sk] = running_agent
result = await runner._handle_message(_make_event("/steer also check auth.log"))
# The handler replied with a confirmation
assert result is not None
assert "steer" in result.lower() or "queued" in result.lower()
# The agent's steer() was called with the payload (prefix stripped)
running_agent.steer.assert_called_once()
injected = running_agent.steer.call_args.args[0]
assert injected.endswith("\n\nalso check auth.log")
assert '"chat_id": "c1"' in injected
# Critically: interrupt was NOT called
running_agent.interrupt.assert_not_called()
# And no user-text queueing happened — the steer doesn't go into
# _pending_messages (that would be turn-boundary /queue semantics).
assert runner._pending_messages == {}
assert adapter._pending_messages == {}
@pytest.mark.asyncio
@pytest.mark.parametrize("elapsed", [float("nan"), True, "not-a-number"])
async def test_steer_reaches_ancient_turn_via_fresh_timestamp_fallback(
monkeypatch, elapsed
):
runner, _adapter = _make_runner(_session_entry())
sk = build_session_key(_make_source())
running_agent = MagicMock()
running_agent.steer.return_value = True
running_agent.get_activity_summary.return_value = {
"seconds_since_activity": elapsed,
"last_activity_at": time.time() - 3,
"last_activity_desc": "receiving stream response",
}
runner._running_agents[sk] = running_agent
runner._running_agents_ts[sk] = time.time() - 31_471
monkeypatch.setenv("HERMES_AGENT_TIMEOUT", "1800")
result = await runner._handle_message(_make_event("/steer pause safely"))
running_agent.steer.assert_called_once()
injected = running_agent.steer.call_args.args[0]
assert injected.endswith("\n\npause safely")
assert '"chat_id": "c1"' in injected
assert runner._running_agents[sk] is running_agent
assert result is not None
@pytest.mark.asyncio
async def test_steer_agent_without_steer_method_falls_back():
"""If the running agent somehow lacks the steer() method (older build,
test stub), the handler must not explode — fall back to /queue."""
runner, adapter = _make_runner(_session_entry())
sk = build_session_key(_make_source())
# A bare object that does NOT have steer() — use a spec'd Mock so
# hasattr(agent, "steer") returns False.
running_agent = MagicMock(spec=[])
runner._running_agents[sk] = running_agent
result = await runner._handle_message(
_make_event("/steer fallback", channel_context="[Thread context]\nAlice: earlier request")
)
assert result is not None
# Must mention queueing since steer wasn't available
assert "queued" in result.lower()
assert sk in adapter._pending_messages
assert adapter._pending_messages[sk].text == "fallback"
assert (
adapter._pending_messages[sk].channel_context
== "[Thread context]\nAlice: earlier request"
)
if __name__ == "__main__": # pragma: no cover
pytest.main([__file__, "-v"])