mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-30 19:09:28 +00:00
Second, deeper pass over tools/gateway/hermes_cli plus first pass over the trees wave 1 missed (acp, acp_adapter, skills, computer_use, docker, dashboard, conformance, monitoring, secret_sources, hermes_state, providers). Same rubric as wave 1 (AGENTS.md test policy); security, alternation/caching invariants, issue-number regressions, and E2E kept. Real test-quality fixes found and rooted out along the way: - tests/tools/test_command_guards.py made real auxiliary-LLM HTTPS calls (DEFAULT_CONFIG smart-approval leaked in) — pinned approval mode=manual via autouse fixture: 17.4s → 0.4s. - test_model_switch_custom_providers.py / test_user_providers_model_switch.py silently probed live provider catalogs (~2s/test) — stubbed cached_provider_model_ids/provider_model_ids/fetch_api_models. - test_telegram_noise_filter.py: 15-platform copy-paste matrix over shared gateway.run logic → 3 representative platforms (55s → 3.9s). - test_gateway_shutdown.py: stop()'s 5s interrupt-deadline loop spun on MagicMock agents — interrupt.side_effect now clears _running_agents (22s → 1.0s). - test_gateway_inactivity_timeout.py poll-harness timings shrunk 3-5x (24s → 1.1s); test_mcp_stability.py backoff/SIGTERM-grace sleeps patched (15.4s → 2.5s); test_async_delegation.py negative-drain wait 5s → 0.5s. - test_telegram_init_deadline.py: loop-block margin restored to 1.0s with rationale comment — the watchdog-dump assertion needs the loop blocked well past deadline+grace under parallel load (flaked once in the 40-worker verification run at a 0.2s margin). Verification: full hermetic suite via scripts/run_tests.sh — 2,438 files, 21,718 tests passed, 0 failed, 293.9s wall. Suite totals vs original baseline: 46,820 → 19,757 test functions (−57.8%), wall 583.5s → 293.9s (−50%), subprocess CPU 13,564s → 11,623s.
297 lines
10 KiB
Python
297 lines
10 KiB
Python
"""Tests for opt-in cleanup of temporary progress bubbles.
|
|
|
|
When ``display.platforms.<plat>.cleanup_progress: true`` is set for a
|
|
platform whose adapter supports message deletion (e.g. Telegram), the
|
|
tool-progress bubble, "⏳ Working — N min" heartbeats, and status-callback
|
|
messages sent during a run are deleted after the final response is
|
|
delivered.
|
|
|
|
Failed runs skip cleanup so the bubbles remain as breadcrumbs.
|
|
Adapters without ``delete_message`` silently no-op.
|
|
"""
|
|
|
|
import asyncio
|
|
import importlib
|
|
import inspect as _inspect
|
|
import sys
|
|
import time
|
|
import types
|
|
from types import SimpleNamespace
|
|
|
|
import pytest
|
|
|
|
from gateway.config import Platform, PlatformConfig
|
|
|
|
|
|
async def _fire_post_delivery_cb(cb):
|
|
"""Invoke a popped post-delivery callback, awaiting if it's async.
|
|
|
|
Chained registrations return an async wrapper; single registrations
|
|
return the raw sync callable. Either way, await any awaitable result.
|
|
"""
|
|
result = cb()
|
|
if _inspect.isawaitable(result):
|
|
await result
|
|
from gateway.platforms.base import BasePlatformAdapter, SendResult
|
|
from gateway.session import SessionSource
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Test fakes — mirror those in test_run_progress_topics.py but add a
|
|
# delete_message implementation that records ids instead of hitting a bot.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
class CleanupCaptureAdapter(BasePlatformAdapter):
|
|
"""Adapter that records every delete_message call for inspection."""
|
|
|
|
_next_mid = 100
|
|
|
|
def __init__(self, platform=Platform.TELEGRAM):
|
|
super().__init__(PlatformConfig(enabled=True, token="***"), platform)
|
|
self.sent = []
|
|
self.edits = []
|
|
self.deleted = []
|
|
|
|
async def connect(self, *, is_reconnect: bool = False) -> bool:
|
|
return True
|
|
|
|
async def disconnect(self) -> None:
|
|
return None
|
|
|
|
def _mint_id(self) -> str:
|
|
CleanupCaptureAdapter._next_mid += 1
|
|
return str(CleanupCaptureAdapter._next_mid)
|
|
|
|
async def send(self, chat_id, content, reply_to=None, metadata=None) -> SendResult:
|
|
mid = self._mint_id()
|
|
self.sent.append(
|
|
{"chat_id": chat_id, "content": content, "message_id": mid, "metadata": metadata}
|
|
)
|
|
return SendResult(success=True, message_id=mid)
|
|
|
|
async def edit_message(self, chat_id, message_id, content) -> SendResult:
|
|
self.edits.append({"chat_id": chat_id, "message_id": message_id, "content": content})
|
|
return SendResult(success=True, message_id=message_id)
|
|
|
|
async def delete_message(self, chat_id, message_id) -> bool:
|
|
self.deleted.append({"chat_id": chat_id, "message_id": str(message_id)})
|
|
return True
|
|
|
|
async def send_typing(self, chat_id, metadata=None) -> None:
|
|
return None
|
|
|
|
async def stop_typing(self, chat_id) -> None:
|
|
return None
|
|
|
|
async def get_chat_info(self, chat_id: str):
|
|
return {"id": chat_id}
|
|
|
|
|
|
class NoDeleteAdapter(CleanupCaptureAdapter):
|
|
"""Adapter that inherits the base no-op delete_message (used to prove
|
|
the cleanup path skips adapters without deletion support)."""
|
|
|
|
async def delete_message(self, chat_id, message_id) -> bool: # type: ignore[override]
|
|
# Pretend to be an adapter whose platform doesn't support deletion:
|
|
# match the base class behavior exactly. gateway/run.py checks
|
|
# ``type(adapter).delete_message is BasePlatformAdapter.delete_message``
|
|
# to detect this, so we re-assign at class body level below.
|
|
raise AssertionError("should not be called — cleanup must skip this adapter")
|
|
|
|
|
|
# Re-bind so the class's delete_message identity equals the base's.
|
|
NoDeleteAdapter.delete_message = BasePlatformAdapter.delete_message
|
|
|
|
|
|
class ProgressAgent:
|
|
"""Emits two tool-progress events and returns a normal final response."""
|
|
|
|
def __init__(self, **kwargs):
|
|
self.tool_progress_callback = kwargs.get("tool_progress_callback")
|
|
self.tools = []
|
|
|
|
def run_conversation(self, message, conversation_history=None, task_id=None):
|
|
cb = self.tool_progress_callback
|
|
if cb is not None:
|
|
cb("tool.started", "terminal", "pwd", {})
|
|
time.sleep(0.2)
|
|
cb("tool.started", "terminal", "ls", {})
|
|
time.sleep(0.2)
|
|
return {"final_response": "done", "messages": [], "api_calls": 1}
|
|
|
|
|
|
class FailingAgent:
|
|
def __init__(self, **kwargs):
|
|
self.tool_progress_callback = kwargs.get("tool_progress_callback")
|
|
self.tools = []
|
|
|
|
def run_conversation(self, message, conversation_history=None, task_id=None):
|
|
cb = self.tool_progress_callback
|
|
if cb is not None:
|
|
cb("tool.started", "terminal", "pwd", {})
|
|
time.sleep(0.2)
|
|
# Empty final_response + failed=True is the shape the gateway
|
|
# actually returns on provider errors (see gateway/run.py where
|
|
# failed keys are only propagated when final_response is empty).
|
|
return {
|
|
"final_response": "",
|
|
"messages": [],
|
|
"api_calls": 1,
|
|
"failed": True,
|
|
"error": "simulated provider failure",
|
|
}
|
|
|
|
|
|
def _make_runner(adapter):
|
|
gateway_run = importlib.import_module("gateway.run")
|
|
GatewayRunner = gateway_run.GatewayRunner
|
|
runner = object.__new__(GatewayRunner)
|
|
runner.adapters = {adapter.platform: adapter}
|
|
runner._voice_mode = {}
|
|
runner._prefill_messages = []
|
|
runner._ephemeral_system_prompt = ""
|
|
runner._reasoning_config = None
|
|
runner._provider_routing = {}
|
|
runner._fallback_model = None
|
|
runner._session_db = None
|
|
runner._running_agents = {}
|
|
runner._session_run_generation = {}
|
|
runner.hooks = SimpleNamespace(loaded_hooks=False)
|
|
runner.config = SimpleNamespace(
|
|
thread_sessions_per_user=False,
|
|
group_sessions_per_user=False,
|
|
stt_enabled=False,
|
|
)
|
|
return runner
|
|
|
|
|
|
def _install_fakes(
|
|
monkeypatch,
|
|
agent_cls,
|
|
*,
|
|
cleanup_on: bool,
|
|
cleanup_platform: Platform = Platform.TELEGRAM,
|
|
):
|
|
"""Wire up the module stubs every _run_agent test needs."""
|
|
monkeypatch.setenv("HERMES_TOOL_PROGRESS_MODE", "all")
|
|
|
|
fake_dotenv = types.ModuleType("dotenv")
|
|
fake_dotenv.load_dotenv = lambda *a, **k: None
|
|
monkeypatch.setitem(sys.modules, "dotenv", fake_dotenv)
|
|
|
|
fake_run_agent = types.ModuleType("run_agent")
|
|
fake_run_agent.AIAgent = agent_cls
|
|
monkeypatch.setitem(sys.modules, "run_agent", fake_run_agent)
|
|
import tools.terminal_tool # noqa: F401 — register tool emoji
|
|
|
|
gateway_run = importlib.import_module("gateway.run")
|
|
monkeypatch.setattr(gateway_run, "_resolve_runtime_agent_kwargs", lambda: {"api_key": "fake"})
|
|
|
|
# Wire the per-platform cleanup_progress flag via the config loader the
|
|
# gateway actually reads (``_load_gateway_config`` returns user config).
|
|
cfg = {
|
|
"display": {
|
|
"platforms": {
|
|
cleanup_platform.value: {"cleanup_progress": True},
|
|
}
|
|
}
|
|
} if cleanup_on else {}
|
|
monkeypatch.setattr(gateway_run, "_load_gateway_config", lambda: cfg)
|
|
return gateway_run
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Tests
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_messaging_agent_forwards_checkpoint_config(monkeypatch, tmp_path):
|
|
"""Writable gateway agents must receive the configured checkpoint limits."""
|
|
captured = {}
|
|
|
|
class CheckpointCaptureAgent(ProgressAgent):
|
|
def __init__(self, **kwargs):
|
|
captured.update(kwargs)
|
|
super().__init__(**kwargs)
|
|
|
|
adapter = CleanupCaptureAdapter()
|
|
runner = _make_runner(adapter)
|
|
gateway_run = _install_fakes(
|
|
monkeypatch, CheckpointCaptureAgent, cleanup_on=False,
|
|
)
|
|
monkeypatch.setattr(gateway_run, "_hermes_home", tmp_path)
|
|
monkeypatch.setattr(
|
|
gateway_run,
|
|
"_load_gateway_config",
|
|
lambda: {
|
|
"checkpoints": {
|
|
"enabled": True,
|
|
"max_snapshots": 9,
|
|
"max_total_size_mb": 444,
|
|
"max_file_size_mb": 6,
|
|
}
|
|
},
|
|
)
|
|
|
|
source = SessionSource(platform=Platform.TELEGRAM, chat_id="-1001")
|
|
result = await runner._run_agent(
|
|
message="hello",
|
|
context_prompt="",
|
|
history=[],
|
|
source=source,
|
|
session_id="sess-checkpoints",
|
|
session_key="agent:main:telegram:group:-1001",
|
|
)
|
|
|
|
assert result["final_response"] == "done"
|
|
assert captured["checkpoints_enabled"] is True
|
|
assert captured["checkpoint_max_snapshots"] == 9
|
|
assert captured["checkpoint_max_total_size_mb"] == 444
|
|
assert captured["checkpoint_max_file_size_mb"] == 6
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_cleanup_chains_with_existing_callback(monkeypatch, tmp_path):
|
|
"""When a bg-review-style callback is already registered, the cleanup
|
|
callback chains with it — both fire, neither clobbers the other."""
|
|
adapter = CleanupCaptureAdapter()
|
|
runner = _make_runner(adapter)
|
|
gateway_run = _install_fakes(monkeypatch, ProgressAgent, cleanup_on=True)
|
|
monkeypatch.setattr(gateway_run, "_hermes_home", tmp_path)
|
|
|
|
source = SessionSource(platform=Platform.TELEGRAM, chat_id="-1001")
|
|
session_key = "agent:main:telegram:group:-1001"
|
|
|
|
pre_existing_fired = []
|
|
|
|
def _preexisting_callback() -> None:
|
|
pre_existing_fired.append(True)
|
|
|
|
# Pre-register a callback with the same generation the run will use
|
|
# (run_generation=None in this test path — matches the default slot).
|
|
adapter.register_post_delivery_callback(session_key, _preexisting_callback)
|
|
|
|
result = await runner._run_agent(
|
|
message="hello",
|
|
context_prompt="",
|
|
history=[],
|
|
source=source,
|
|
session_id="sess-1",
|
|
session_key=session_key,
|
|
)
|
|
|
|
assert result["final_response"] == "done"
|
|
cb = adapter.pop_post_delivery_callback(session_key)
|
|
assert callable(cb)
|
|
await _fire_post_delivery_cb(cb)
|
|
for _ in range(20):
|
|
await asyncio.sleep(0.01)
|
|
if adapter.deleted:
|
|
break
|
|
|
|
# Both effects land: the pre-existing callback fires AND the cleanup
|
|
# deletes at least one progress bubble.
|
|
assert pre_existing_fired == [True]
|
|
assert len(adapter.deleted) >= 1
|