## Summary
Reported as "deliver_files is broken". It is not: **the tool never ran.**
Every `deliver_files` call on one install was refused before dispatch -- five in a row in a single session -- and nothing was ever delivered:
```
Error: Invalid parameters for tool 'deliver_files': files should be array
```
What the model sent was the array, spelled as a string:
```json
{"files": "[{\"path\": \"...poster.png\", \"title\": \"...\"}]", "message": "..."}
```
`cast_params` already forgives that spelling for every scalar type -- `"3"` for an integer, `"true"` for a boolean -- and did not for arrays or objects, which are exactly the types a model is most likely to stringify.
## Type
- [x] Fix
- [ ] Feature
## The edge of the forgiving is the design
When the value is a string, try `json.loads`, and take the result **only if it is the type the schema asked for**. A string that is not JSON, or is JSON of the wrong shape, is returned unchanged so the validator still names the real problem -- the point is to widen what is accepted, not to swallow a type error.
**Nothing here can raise.** Turning "your argument has the wrong type" into a traceback would be worse than the refusal it replaces, so the parse is wrapped and failure falls through.
**A `type: "string"` field is never parsed.** JSON is ordinary text in a message body or a file's content, and replacing it would hand the tool something it never asked for. Two independent things prevent it today (the string branch returns before the parse is reached, and the parse only wins on a type match), which is why the case pinning it is a tripwire rather than a load-bearing assertion -- said so in its own docstring rather than left to look like coverage.
## Not just deliver_files
Six built-in tools declare an array or object parameter and were all exposed:
| tool | parameter |
|---|---|
| `deliver_files` | `files` |
| `ask_user` | `questions` |
| `image_generate` | `images` |
| `message` | `media` |
| `create_playbook` | `skills` |
| `video_generate` | `params` |
The same failure was already seen from the other side: `acp_client/acp_agent.py` records `update_task_state` and `operations should be array` from a real dispatch.
## Verification
```
make lint-python ruff check + ruff format --check, 1850 files
make lint-imports 9 contracts kept, 0 broken
pytest tests/ 19,699 passed, 7 failed (all pre-existing, see below)
```
- [x] Relevant tests pass locally
- [x] Relevant lint / type checks pass locally
- [ ] User-facing docs or screenshots -- no user-facing string
**End to end, through the real tool**, with the exact payload that failed five times in production:
```
files sent as: str -> [{"path": ".../poster.png", "title": "..."}]
returned: Delivered 1 file(s): poster.png (72 B).
still "should be array": False
registered in the store: True
```
Twelve cases, five mutations, each killing at least one:
| mutation | red |
|---|---|
| drop the whole branch (the old behaviour) | 3 |
| raise instead of falling through on a parse failure | 5 |
| accept the parse whatever type it yielded | 1 |
| do not re-cast the parsed value, so items are not cast | 1 |
| widen the tuple to include `string` | **survived -- unreachable** |
The last is worth stating rather than hiding: that mutation cannot change behaviour, because the `type == "string"` branch returns earlier and the type check would refuse the result anyway. I also tried moving the new block above that branch, alone and combined with widening the tuple, and neither reddens anything either. The invariant holds by construction; the case that documents it says so.
**Pre-existing failures, checked against the base.** 7 failed in the full run, all in `tests/test_security_network.py` (`test_a_redirect_to_a_dangerous_address_is_refused` and siblings). Running that file alone on this same tree gives 4 -- the count is order-dependent -- and the change touches only `raven/agent/tools/params.py` plus a new test file, neither of which `raven/security/network.py` imports. The failures are the box's resolver: it returns poisoned answers roughly half the time, which I measured separately (`en.wikipedia.org` came back as `2001::1` on 3 of 6 lookups). Not introduced here, and worth someone's attention on its own.
## Risk
- [x] Security impact considered -- widens what reaches a tool's own validation, not what bypasses it. A parsed value is still validated against the same schema, and a parse that yields the wrong type is refused exactly as before
- [x] Backward compatibility considered -- a call that worked before takes the identical path; the new branch is reached only by a `str` where an array or object is declared, which previously always failed
- [x] Rollback path is clear -- single-commit revert
## Related Issues
N/A
## Swept
`mr-review-patterns` pre-submit sweep, over `git diff origin/main...HEAD`.
- **1 (a write location moved)** -- nothing moved; a branch was added to an existing dispatch on `target_type`. Grepped every caller of `cast_params`: the registry is the only one, and it runs `validate_params` right after, which is the reader this affects.
- **2 (a value with several entry points)** -- the cast is the single funnel every tool argument passes through, which is why the fix belongs here and not in `deliver_files`.
- **3 (a guard)** -- checked both directions: what it wrongly lets through (JSON of the wrong type, non-JSON text -- both refused, with cases) and what it wrongly refuses (a real list, unaffected, with a case).
- **4 (tests)** -- five mutations above, one honestly reported as unreachable. Failing-ID sets diffed against the base rather than counts.
- **6 (stale text)** -- the module docstring enumerated what is coerced ("a string that should be an integer"); it now names arrays and objects too. Grepped the tracked tree for other statements of the rule: none.
- **9 (error handling)** -- the new path adds one way to end and it is caught: `json.loads` raises `ValueError` on bad text and `TypeError` on a non-string, and both fall through to the untouched value.
- **11** -- `git ls-files | grep -c node_modules` is 0; the commit-message gate was run, not eyeballed.
Not applicable, with reasons: **3b, 3c, 3d, 3e** (no tool schema produced, no state moved, one lane, no assertion lost its symbol); **4b, 4c, 4d** (no platform branch, no shell, no rendered number); **5, 5b** (no dependency; `json` is stdlib and already imported across this package); **6b, 6c** (no backup, no generated artifact); **7, 7b** (nothing on disk); **8** (no asset, no secret); **10** (no prompt text).
## Deliberately not fixed
The error text still ends with `[Analyze the error above and try a different approach.]`, and the model plainly did not -- it sent the same shape five times. A message naming the expected spelling (`files must be a JSON array, not a string containing one`) would give it something to act on. That is a change to the registry's error surface, affects every tool, and deserves its own argument.
## Moved from !594
Opened on a compliant source branch. AGENTS.md 2.1 asks for a 3-5 word snake_case description and the previous branch had seven; GitLab cannot change an existing merge request's source branch, so the change moves rather than the name.
Same commits, rebased onto `c2a9768f`. Both code findings from the review on !594 are already in them -- `RecursionError` caught with its regression case, and the implementation comment reduced to the two non-obvious reasons. Nothing else changed, so that review carries.
539 lines
17 KiB
Python
539 lines
17 KiB
Python
"""Unit tests for TaskDiscoverer + dispatch_options + render_menu_markdown (MS2)."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import json
|
|
from datetime import datetime
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
from raven.memory_engine.consolidate.consolidator import MemoryStore
|
|
from raven.proactive_engine.sentinel.executor.dispatcher import (
|
|
NudgeDispatcher,
|
|
render_menu_markdown,
|
|
)
|
|
from raven.proactive_engine.sentinel.executor.pending_decision import PendingDecisionStore
|
|
from raven.proactive_engine.sentinel.predictor.task_discoverer import (
|
|
TaskDiscoverer,
|
|
_extract_sentinel_observations_block,
|
|
)
|
|
from raven.proactive_engine.sentinel.types import PendingDecision, TaskOption
|
|
|
|
_NOW = datetime(2026, 5, 8, 8, 0)
|
|
_NOW_MS = int(_NOW.timestamp() * 1000)
|
|
|
|
|
|
def _wire_dispatcher(now_fn):
|
|
posted: list = []
|
|
d = NudgeDispatcher(now_fn=now_fn)
|
|
|
|
async def _post(out):
|
|
posted.append(out)
|
|
|
|
d.set_post(_post)
|
|
return d, posted
|
|
|
|
|
|
# ── helpers ───────────────────────────────────────────────────────────
|
|
|
|
|
|
class _StubResponse:
|
|
"""Mimics LLMProvider.chat_with_retry's response object enough for
|
|
the parts TaskDiscoverer touches (has_tool_calls + tool_calls list
|
|
of objects with `.arguments`)."""
|
|
|
|
def __init__(self, options: list[dict] | None, *, raw_args=None):
|
|
if raw_args is not None:
|
|
self._args = raw_args
|
|
elif options is None:
|
|
self._args = None
|
|
else:
|
|
self._args = json.dumps({"options": options})
|
|
self.has_tool_calls = self._args is not None
|
|
if self.has_tool_calls:
|
|
|
|
class _Call:
|
|
arguments = self._args
|
|
|
|
_Call.arguments = self._args
|
|
self.tool_calls = [_Call()]
|
|
else:
|
|
self.tool_calls = []
|
|
|
|
|
|
class _StubProvider:
|
|
"""Async LLM provider stub with configurable canned response."""
|
|
|
|
def __init__(self, response):
|
|
self._response = response
|
|
self.calls: list[dict] = []
|
|
|
|
async def chat_with_retry(self, *, messages, tools, model, tool_choice):
|
|
self.calls.append(
|
|
{
|
|
"messages": messages,
|
|
"tools": tools,
|
|
"model": model,
|
|
"tool_choice": tool_choice,
|
|
}
|
|
)
|
|
return self._response
|
|
|
|
|
|
def _option_dict(**overrides):
|
|
base = {
|
|
"title": "草拟回复 X",
|
|
"why": "X 在昨天发了你尚未回复",
|
|
"type": "ad_hoc",
|
|
"exec_kind": "reply",
|
|
"exec_payload": {"prompt": "请帮我草拟回复 X"},
|
|
"source": "history",
|
|
"priority": "medium",
|
|
}
|
|
base.update(overrides)
|
|
return base
|
|
|
|
|
|
@pytest.fixture
|
|
def workspace(tmp_path: Path) -> Path:
|
|
ws = tmp_path / "chanwork"
|
|
ws.mkdir()
|
|
(ws / "memory").mkdir()
|
|
return ws
|
|
|
|
|
|
@pytest.fixture
|
|
def memory_store(workspace: Path) -> MemoryStore:
|
|
store = MemoryStore(workspace)
|
|
# Plant some content so discovery prompt has substance to work with
|
|
store.write_long_term(
|
|
"## User Information\n"
|
|
"- name: Alice\n"
|
|
"\n"
|
|
"## Sentinel Observations (auto)\n"
|
|
"<!-- sentinel:auto last_updated=2026-05-07T08:00 -->\n"
|
|
"\n"
|
|
"### Signal counts (last 7 days)\n"
|
|
"- dispatched: 47, accepted: 36 (76%)\n"
|
|
"\n"
|
|
"<!-- /sentinel:auto -->\n"
|
|
)
|
|
store.append_history("[2026-05-08 06:30] User started morning routine")
|
|
return store
|
|
|
|
|
|
@pytest.fixture
|
|
def pending_store(tmp_path: Path) -> PendingDecisionStore:
|
|
return PendingDecisionStore(tmp_path / "pending.json")
|
|
|
|
|
|
# ── render_menu_markdown ──────────────────────────────────────────────
|
|
|
|
|
|
def test_render_menu_markdown_format():
|
|
decision = PendingDecision(
|
|
decision_id="dec_x",
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
created_at_ms=_NOW_MS,
|
|
ttl_min=60,
|
|
options=[
|
|
TaskOption(
|
|
id="opt_a",
|
|
title="草拟回复 X",
|
|
why="X 在昨天发了未回复",
|
|
type="ad_hoc",
|
|
exec_kind="reply",
|
|
exec_payload={},
|
|
),
|
|
TaskOption(
|
|
id="opt_b",
|
|
title="周二 PR review",
|
|
why="最近 3 周这么做过",
|
|
type="routine_confirm",
|
|
exec_kind="routine_confirm",
|
|
exec_payload={"routine_id": "dow1-h09-pr"},
|
|
),
|
|
],
|
|
)
|
|
text = render_menu_markdown(decision)
|
|
assert text.startswith("📋 [Today's suggestions]")
|
|
assert text.splitlines()[2].startswith("1. (new task) ")
|
|
assert " — X 在昨天发了未回复" in text
|
|
assert "2. (ongoing ✓) " in text
|
|
assert "Reply with a number to choose" in text
|
|
|
|
|
|
def test_render_menu_markdown_omits_why_when_empty():
|
|
decision = PendingDecision(
|
|
decision_id="dec_x",
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
created_at_ms=_NOW_MS,
|
|
options=[
|
|
TaskOption(id="opt_a", title="bare option", why="", type="ad_hoc", exec_kind="reply", exec_payload={}),
|
|
],
|
|
)
|
|
text = render_menu_markdown(decision)
|
|
assert "1. (new task) bare option" in text
|
|
# No "—" line for empty why
|
|
assert " —" not in text.split("回复数字")[0]
|
|
|
|
|
|
# ── _extract_sentinel_observations_block ──────────────────────────────
|
|
|
|
|
|
def test_extract_observations_block_pulls_section_body():
|
|
md = (
|
|
"## User Information\n"
|
|
"- foo\n"
|
|
"\n"
|
|
"## Sentinel Observations (auto)\n"
|
|
"<!-- sentinel:auto last_updated=2026-05-07T08:00 -->\n"
|
|
"BODY HERE\n"
|
|
"<!-- /sentinel:auto -->\n"
|
|
"\n"
|
|
"## Other\n"
|
|
)
|
|
body = _extract_sentinel_observations_block(md)
|
|
assert body == "BODY HERE"
|
|
|
|
|
|
def test_extract_observations_block_returns_empty_when_missing():
|
|
assert _extract_sentinel_observations_block("") == ""
|
|
assert _extract_sentinel_observations_block("## Foo\nbar") == ""
|
|
|
|
|
|
# ── TaskDiscoverer.run happy path ─────────────────────────────────────
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_supersede_notice_submits_sentinel_origin_when_wired(memory_store, pending_store):
|
|
# Spine path: with submit wired, the supersede notice runs as a
|
|
# SENTINEL-origin turn (a system notice — after_send is skipped) instead of
|
|
# publishing to the bus.
|
|
from raven.spine import Origin
|
|
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=_StubProvider(_StubResponse([])),
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
captured = {}
|
|
|
|
class _Handle:
|
|
async def result(self):
|
|
return None
|
|
|
|
disco.set_submit(lambda req: (captured.__setitem__("req", req), _Handle())[1])
|
|
|
|
await disco._notify_superseded_awaiting(
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
superseded_ids=["dec_old"],
|
|
)
|
|
|
|
req = captured["req"]
|
|
assert req.origin is Origin.SENTINEL
|
|
assert req.source.sender_id == "sentinel"
|
|
assert req.source.channel == "feishu" and req.source.chat_id == "ou_xxx"
|
|
assert "replaced" in req.text
|
|
assert req.sentinel is None # not a menu-pick → no action_origin
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_happy_path_creates_decision_and_dispatches(memory_store, pending_store):
|
|
dispatcher, posted = _wire_dispatcher(lambda: _NOW)
|
|
response = _StubResponse(
|
|
[
|
|
_option_dict(title="task A"),
|
|
_option_dict(title="task B", priority="high"),
|
|
_option_dict(title="task C", source="memory"),
|
|
]
|
|
)
|
|
provider = _StubProvider(response)
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="qwen3.5-27B",
|
|
max_options=4,
|
|
decision_ttl_min=60,
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert decision is not None
|
|
assert len(decision.options) == 3
|
|
titles = [o.title for o in decision.options]
|
|
assert titles == ["task A", "task B", "task C"]
|
|
|
|
# Persisted
|
|
fetched = pending_store.get_recent("feishu", "ou_xxx", now_ms=_NOW_MS + 1)
|
|
assert fetched is not None
|
|
assert fetched.decision_id == decision.decision_id
|
|
|
|
outbound = posted.pop(0)
|
|
assert outbound.source.channel == "feishu"
|
|
assert outbound.source.chat_id == "ou_xxx"
|
|
assert outbound.source.extras["_sentinel_action"] == "discovery_menu"
|
|
assert "task A" in outbound.content
|
|
assert "task B" in outbound.content
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_returns_none_on_no_tool_call(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
provider = _StubProvider(_StubResponse(options=None))
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
result = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert result is None
|
|
# Nothing persisted
|
|
assert pending_store.get_recent("feishu", "ou_xxx", now_ms=_NOW_MS + 1) is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_drops_malformed_options_keeps_valid(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
response = _StubResponse(
|
|
[
|
|
_option_dict(title="valid one"),
|
|
{"title": "missing fields"}, # missing exec_kind+payload → invalid
|
|
_option_dict(title="", exec_kind="reply"), # empty title → dropped
|
|
_option_dict(title="bad kind", exec_kind="weird_kind"), # unknown kind
|
|
_option_dict(title="another valid", priority="high"),
|
|
]
|
|
)
|
|
provider = _StubProvider(response)
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert decision is not None
|
|
titles = [o.title for o in decision.options]
|
|
assert titles == ["valid one", "another valid"]
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_rejects_routine_confirm_without_routine_id(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
# pad with 2 ad_hoc options so we get the minimum 3
|
|
response = _StubResponse(
|
|
[
|
|
_option_dict(
|
|
title="bad routine",
|
|
type="routine_confirm",
|
|
exec_kind="routine_confirm",
|
|
exec_payload={},
|
|
),
|
|
_option_dict(
|
|
title="good routine",
|
|
type="routine_confirm",
|
|
exec_kind="routine_confirm",
|
|
exec_payload={"routine_id": "dow1-h09-meeting"},
|
|
),
|
|
_option_dict(title="ad hoc 1"),
|
|
_option_dict(title="ad hoc 2"),
|
|
]
|
|
)
|
|
provider = _StubProvider(response)
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert decision is not None
|
|
titles = [o.title for o in decision.options]
|
|
assert "bad routine" not in titles
|
|
assert "good routine" in titles
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_truncates_to_max_options(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
response = _StubResponse([_option_dict(title=f"option {i}") for i in range(8)])
|
|
provider = _StubProvider(response)
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
max_options=3, # keep only the first 3
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert decision is not None
|
|
assert len(decision.options) == 3
|
|
titles = [o.title for o in decision.options]
|
|
assert titles == ["option 0", "option 1", "option 2"]
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_overdue_survives_truncation(memory_store, pending_store):
|
|
"""Regression: an overdue option emitted past max_options must float to
|
|
the front BEFORE truncation, not get dropped (annotate-then-truncate)."""
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
opts = [_option_dict(title=f"option {i}") for i in range(5)]
|
|
# Overdue task emitted LAST (index 5), deadline before _NOW (2026-05-08).
|
|
opts.append(_option_dict(title="交月报", deadline="2026-05-01"))
|
|
provider = _StubProvider(_StubResponse(opts))
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
max_options=3,
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert decision is not None
|
|
assert len(decision.options) == 3
|
|
# Floated to the front with the overdue marker; survived truncation.
|
|
assert decision.options[0].title.startswith("⚠️ overdue 5/1 ")
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_handles_malformed_json_args_gracefully(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
# Hand-craft response with non-JSON arg string
|
|
response = _StubResponse(options=None, raw_args="this is not json")
|
|
response.has_tool_calls = True
|
|
|
|
class _C:
|
|
arguments = "this is not json"
|
|
|
|
response.tool_calls = [_C()]
|
|
provider = _StubProvider(response)
|
|
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
result = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert result is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_discoverer_supersedes_prior_decision_on_same_address(memory_store, pending_store):
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
|
|
# Plant a pre-existing decision
|
|
pre_existing = PendingDecision(
|
|
decision_id="dec_old",
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
created_at_ms=_NOW_MS - 1000,
|
|
options=[TaskOption(id="opt_old", title="old", why="", type="ad_hoc", exec_kind="reply", exec_payload={})],
|
|
)
|
|
pending_store.put(pre_existing)
|
|
|
|
response = _StubResponse(
|
|
[
|
|
_option_dict(title="new task A"),
|
|
_option_dict(title="new task B"),
|
|
_option_dict(title="new task C"),
|
|
]
|
|
)
|
|
provider = _StubProvider(response)
|
|
disco = TaskDiscoverer(
|
|
memory_store=memory_store,
|
|
pending_store=pending_store,
|
|
dispatcher=dispatcher,
|
|
provider=provider,
|
|
model="x",
|
|
now_fn=lambda: _NOW,
|
|
)
|
|
|
|
new_decision = await disco.run(channel="feishu", to="ou_xxx")
|
|
assert new_decision is not None
|
|
|
|
fetched = pending_store.get_recent("feishu", "ou_xxx", now_ms=_NOW_MS + 1)
|
|
assert fetched is not None
|
|
assert fetched.decision_id == new_decision.decision_id # not "dec_old"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_dispatcher_dispatch_options_publishes_outbound(memory_store, pending_store):
|
|
"""Discovery menus go to OUTBOUND so the user sees render_menu_markdown
|
|
verbatim (no agent paraphrase). DecisionConsumerAdapter handles the
|
|
pick path independently — see dispatcher.dispatch_options docstring."""
|
|
dispatcher, posted = _wire_dispatcher(lambda: _NOW)
|
|
decision = PendingDecision(
|
|
decision_id="dec_x",
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
created_at_ms=_NOW_MS,
|
|
options=[
|
|
TaskOption(
|
|
id="opt_a", title="alpha", why="why-a", type="ad_hoc", exec_kind="reply", exec_payload={"prompt": "go"}
|
|
)
|
|
],
|
|
)
|
|
|
|
result = await dispatcher.dispatch_options(decision)
|
|
assert result.delivered is True
|
|
assert result.details["decision_id"] == "dec_x"
|
|
|
|
outbound = posted.pop(0)
|
|
assert outbound.source.channel == "feishu"
|
|
assert outbound.source.chat_id == "ou_xxx"
|
|
assert outbound.source.extras["_sentinel_origin"] is True
|
|
assert outbound.source.extras["_sentinel_decision_id"] == "dec_x"
|
|
assert "alpha" in outbound.content
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_dispatcher_dispatch_options_rejects_empty():
|
|
dispatcher, _posted = _wire_dispatcher(lambda: _NOW)
|
|
decision = PendingDecision(
|
|
decision_id="dec_x",
|
|
channel="feishu",
|
|
to="ou_xxx",
|
|
created_at_ms=_NOW_MS,
|
|
options=[],
|
|
)
|
|
result = await dispatcher.dispatch_options(decision)
|
|
assert result.delivered is False
|
|
assert result.reason == "empty_options"
|