hermes-agent/tests/tui_gateway/test_todo_state_events.py

213 lines
9.6 KiB
Python

"""Todo snapshots bypass optional tool-progress display settings."""
import json
from hermes_state import SessionDB
import tui_gateway.server as server
def test_deferred_desktop_resume_exposes_todo_older_than_display_tail(tmp_path, monkeypatch):
db = SessionDB(tmp_path / "state.db")
state = {"todos": [{"id": "a", "content": "Older plan", "status": "in_progress"}], "revision": 4}
db.create_session("todo-parent", source="desktop")
db.append_message("todo-parent", "assistant", "", tool_calls=[
{"id": "todo-call", "type": "function", "function": {
"name": "tool_call", "arguments": json.dumps({"calls": [{"name": "todo_list", "arguments": {}}]}),
}},
])
db.append_message("todo-parent", "tool", json.dumps(state), tool_call_id="todo-call", tool_name="todo_list")
db.end_session("todo-parent", "compression")
db.create_session("todo-tip", source="desktop", parent_session_id="todo-parent")
for i in range(130):
db.append_message("todo-tip", "user" if i % 2 == 0 else "assistant", f"Later row {i}")
monkeypatch.setattr(server, "_get_db", lambda: db)
monkeypatch.setattr(server, "_default_session_cwd", lambda: str(tmp_path))
monkeypatch.setattr(server, "_profile_configured_cwd", lambda _: str(tmp_path))
monkeypatch.setattr(server, "_enable_gateway_prompts", lambda: None)
monkeypatch.setattr(server, "_schedule_resume_hydration", lambda *args, **kwargs: None)
monkeypatch.setattr(server, "_schedule_session_cap_enforcement", lambda: None)
sid = None
try:
assert json.loads(db.get_latest_todo_result("todo-tip")) == state
response = server.handle_request({"id": "resume-todo", "method": "session.resume", "params": {
"session_id": "todo-tip", "source": "desktop", "defer_history": True, "omit_messages": True,
}})
assert "error" not in response, response
sid = response["result"]["session_id"]
assert response["result"]["todo_state"] == state
assert response["result"]["messages"] == []
# The latest explicit clear wins over an older ancestor's plan.
empty = {"todos": [], "revision": 5}
db.append_message("todo-tip", "user", "Clear the old plan")
db.append_message("todo-tip", "assistant", "", tool_calls=[
{"id": "clear-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("todo-tip", "tool", json.dumps(empty), tool_call_id="clear-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("todo-tip")) == empty
# A long ancestor history must not hide the child's explicit clear.
db.create_session("long-parent", source="desktop")
for i in range(180):
db.append_message("long-parent", "user", f"Parent row {i}")
db.append_message("long-parent", "assistant", "", tool_calls=[
{"id": "parent-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("long-parent", "tool", json.dumps(state), tool_call_id="parent-call", tool_name="todo_list")
db.create_session("short-child", source="desktop", parent_session_id="long-parent")
db.append_message("short-child", "assistant", "", tool_calls=[
{"id": "child-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("short-child", "tool", json.dumps(empty), tool_call_id="child-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("short-child")) == empty
# A branch is a copied transcript, not a live view of its parent.
db.create_session("independent-branch", source="desktop", parent_session_id="long-parent",
model_config={"_branched_from": "long-parent"})
db.append_message("independent-branch", "assistant", "", tool_calls=[
{"id": "branch-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("independent-branch", "tool", json.dumps(state), tool_call_id="branch-call", tool_name="todo_list")
db.append_message("long-parent", "assistant", "", tool_calls=[
{"id": "parent-new-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("long-parent", "tool", json.dumps(empty), tool_call_id="parent-new-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("independent-branch")) == state
# Unpaired tool rows must not seed a resumed checklist, even if well-formed.
db.append_message("independent-branch", "user", "An unrelated turn")
db.append_message("independent-branch", "tool", json.dumps(empty), tool_call_id="forged-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("independent-branch")) == state
# Malformed/oversized latest rows must not conceal an earlier valid plan.
db.append_message("independent-branch", "assistant", "", tool_calls=[
{"id": "malformed-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("independent-branch", "tool", "{not valid json}", tool_call_id="malformed-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("independent-branch")) == state
oversized = json.dumps({**empty, "padding": "x" * 512_001})
db.append_message("independent-branch", "assistant", "", tool_calls=[
{"id": "oversized-call", "type": "function", "function": {"name": "todo_list", "arguments": "{}"}},
])
db.append_message("independent-branch", "tool", oversized, tool_call_id="oversized-call", tool_name="todo_list")
assert json.loads(db.get_latest_todo_result("independent-branch")) == state
finally:
if sid:
server._sessions.pop(sid, None)
db.close()
def test_todo_completion_always_emits_snapshot_and_compat_event(monkeypatch):
sid = "todo-state-test"
events = []
session = {
"agent": None,
"edit_snapshots": {},
"tool_started_at": {},
"tool_progress_mode": "off",
}
monkeypatch.setitem(server._sessions, sid, session)
monkeypatch.setattr(server, "_tool_progress_enabled", lambda _sid: False)
monkeypatch.setattr(server, "_tool_lifecycle_required_for_ui", lambda _name: False)
monkeypatch.setattr(
server,
"_emit",
lambda event, event_sid, payload=None: events.append(
(event, event_sid, payload)
),
)
state = {
"todos": [{"id": "1", "content": "Work", "status": "in_progress"}],
"revision": 9,
}
server._on_tool_complete(sid, "call-1", "todo", {}, json.dumps(state))
assert [event[0] for event in events] == ["tool.complete", "todo.updated"]
assert events[-1] == ("todo.updated", sid, state)
assert session["todo_state"] == state
def test_non_todo_completion_stays_suppressed_when_progress_is_off(monkeypatch):
sid = "ordinary-tool-test"
events = []
monkeypatch.setitem(
server._sessions,
sid,
{
"agent": None,
"edit_snapshots": {},
"tool_started_at": {},
"tool_progress_mode": "off",
},
)
monkeypatch.setattr(server, "_tool_progress_enabled", lambda _sid: False)
monkeypatch.setattr(server, "_tool_lifecycle_required_for_ui", lambda _name: False)
monkeypatch.setattr(server, "_emit", lambda *args: events.append(args))
server._on_tool_complete(sid, "call-1", "terminal", {}, "ok")
assert events == []
def test_live_snapshot_prefers_the_highest_revision():
class Store:
@staticmethod
def snapshot():
return {"todos": [], "revision": 4}
class Agent:
_todo_store = Store()
session = {
"agent": Agent(),
"todo_state": {
"todos": [{"id": "1", "content": "Current", "status": "pending"}],
"revision": 5,
},
}
payload = server._attach_todo_state({}, session)
assert payload["todo_state"]["revision"] == 5
def test_unused_store_is_not_attached():
class Store:
@staticmethod
def snapshot():
return {"todos": [], "revision": 0}
class Agent:
_todo_store = Store()
payload = server._attach_todo_state({}, {"agent": Agent()})
assert "todo_state" not in payload
def test_empty_list_at_nonzero_revision_is_a_real_clear():
state = server._normalize_todo_state({"todos": [], "revision": 2})
assert state == {"todos": [], "revision": 2}
def test_subagent_lifecycle_bypasses_tool_progress_off(monkeypatch):
"""Subagent rows feed the Desktop status stack / TUI spawn tree — application state, not
tool-progress chrome — so display.tool_progress=off must not swallow them."""
sid = "subagent-progress-off"
events = []
monkeypatch.setitem(server._sessions, sid, {"agent": None, "tool_progress_mode": "off"})
monkeypatch.setattr(server, "_tool_progress_enabled", lambda _sid: False)
monkeypatch.setattr(server, "_emit", lambda event, event_sid, payload=None: events.append(event))
server._on_tool_progress(sid, "subagent.start", "delegate_task", "goal", None, goal="goal", subagent_id="s1")
# tool.output_risk is tool chrome, so tool_progress=off drops it. (reasoning.available is
# not a control here: reasoning follows display.show_reasoning, never tool_progress.)
server._on_tool_progress(
sid, "tool.output_risk", "terminal", None, None,
tool_call_id="t1", risk_metadata={"risk": "high", "findings": []},
)
assert events == ["subagent.start"]