Raven-1/tests/test_personalizer.py
Handsome-wzw 7b45527b5b fix(tools): an array sent as its own JSON is still an array
## 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.
2026-09-09 15:42:36 +00:00

173 lines
7.8 KiB
Python

"""The personalizer, driven directly rather than through a turn.
Every method is written to fail neutrally: the agent loop calls these between a
user's message and the model's answer, so a bad LLM reply or a provider outage
has to cost a clarification, never the turn. That is the property these pin --
along with the two shapes the extraction prompt is allowed to answer in, and
what reaches user.md when it answers well.
"""
from __future__ import annotations
from pathlib import Path
from types import SimpleNamespace
import pytest
from raven.agent.personalizer.personalizer import Personalizer
from raven.memory_engine.consolidate.consolidator import MemoryStore
class _Provider:
"""Answers every chat with the next queued reply, or raises the queued error."""
def __init__(self, *replies: str | Exception) -> None:
self._replies = list(replies)
self.calls: list[dict] = []
async def chat(self, **kwargs):
self.calls.append(kwargs)
reply = self._replies.pop(0) if self._replies else ""
if isinstance(reply, Exception):
raise reply
return SimpleNamespace(content=reply)
def _personalizer(tmp_path: Path, *replies: str | Exception) -> Personalizer:
return Personalizer(MemoryStore(tmp_path), _Provider(*replies), "stub-model")
# ── reading a model's answer ────────────────────────────────────────────────
class TestReadingTheModelsAnswer:
def test_json_wrapped_in_prose_is_still_read(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path)
assert p._parse_json('Sure! {"a": 1} hope that helps', {"a": 0}) == {"a": 1}
def test_an_answer_with_no_object_is_the_fallback(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path)
assert p._parse_json("I could not decide.", {"needs_clarification": False}) == {"needs_clarification": False}
def test_a_broken_object_is_the_fallback_not_an_exception(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path)
assert p._parse_json('{"a": }', {"a": 0}) == {"a": 0}
# ── the two shapes a fact list arrives in ───────────────────────────────────
class TestGroupingFacts:
def test_a_category_chooses_the_section(self) -> None:
grouped = Personalizer._group_facts_by_category(
[
{"text": "prefers Go", "category": "preference"},
{"text": "asks before nudging", "category": "proactivity"},
]
)
assert grouped == {
"Preferences": ["prefers Go"],
"Proactivity Preferences": ["asks before nudging"],
}
def test_a_category_nobody_maps_falls_back_to_the_section_the_caller_named(self) -> None:
grouped = Personalizer._group_facts_by_category(
[{"text": "runs on a laptop", "category": "hardware"}], legacy_section="Context"
)
assert grouped == {"Context": ["runs on a laptop"]}
def test_the_legacy_shape_is_a_flat_list_under_one_section(self) -> None:
grouped = Personalizer._group_facts_by_category(["prefers Go", " "], legacy_section="Context")
assert grouped == {"Context": ["prefers Go"]}
def test_items_that_carry_no_text_are_dropped(self) -> None:
assert Personalizer._group_facts_by_category([{"text": " "}, None, 7, []]) == {}
assert Personalizer._group_facts_by_category([]) == {}
assert Personalizer._group_facts_by_category(None) == {}
# ── what the prompt is told about the conversation ──────────────────────────
class TestFormattingHistory:
def test_only_the_last_few_messages_travel(self) -> None:
history = [{"role": "user", "content": f"m{i}"} for i in range(6)]
assert Personalizer._format_history(history) == "USER: m2\nUSER: m3\nUSER: m4\nUSER: m5"
def test_a_long_message_is_truncated(self) -> None:
out = Personalizer._format_history([{"role": "user", "content": "x" * 250}])
assert out == "USER: " + "x" * 200 + "..."
def test_nothing_usable_reads_as_no_prior_context(self) -> None:
assert Personalizer._format_history([{"role": "user", "content": " "}]) == "(no prior context)"
assert Personalizer._format_history([]) == "(no prior context)"
# ── writing to user.md ──────────────────────────────────────────────────────
class TestWritingASection:
def test_facts_land_under_a_header_that_is_already_there(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path)
p.memory.write_long_term("# Me\n\n## Preferences\n- likes tea\n\n## Other\n- keep\n")
p._append_to_memory_section("Preferences", ["prefers Go"])
assert p.memory.read_long_term() == ("# Me\n\n## Preferences\n- prefers Go\n- likes tea\n\n## Other\n- keep\n")
def test_a_missing_header_is_created_at_the_end(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path)
p.memory.write_long_term("# Me\n\n## Other\n- keep\n")
p._append_to_memory_section("Proactivity Preferences", ["asks first", "quiet at night"])
assert p.memory.read_long_term().endswith("## Proactivity Preferences\n- asks first\n- quiet at night\n")
assert "## Other" in p.memory.read_long_term()
# ── the steps the loop calls ────────────────────────────────────────────────
class TestTheStepsTheLoopCalls:
async def test_classify_reads_the_models_verdict(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path, 'Thinking... {"needs_clarification": true, "domain": "code"}')
assert await p.classify("write me a parser") == {"needs_clarification": True, "domain": "code"}
async def test_a_provider_that_raises_costs_the_clarification_not_the_turn(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path, RuntimeError("provider down"))
assert await p.classify("write me a parser") == {"needs_clarification": False, "domain": ""}
async def test_an_answer_with_no_reusable_preference_writes_nothing(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path, '{"facts": [], "section": "Preferences"}')
p.memory.write_long_term("# Me\n")
assert await p.extract_and_store_preference("do it", "which language?", "whatever") is False
assert p.memory.read_long_term() == "# Me\n"
async def test_an_extracted_preference_reaches_the_section_it_named(self, tmp_path: Path) -> None:
p = _personalizer(tmp_path, '{"facts": ["prefers Go"], "section": "Preferences"}')
p.memory.write_long_term("# Me\n")
assert await p.extract_and_store_preference("do it", "which language?", "Go") is True
assert "## Preferences\n- prefers Go" in p.memory.read_long_term()
async def test_the_history_reaches_the_prompt(self, tmp_path: Path) -> None:
provider = _Provider('{"needs_clarification": false, "domain": ""}')
p = Personalizer(MemoryStore(tmp_path), provider, "stub-model")
await p.classify("and now?", history=[{"role": "user", "content": "earlier thing"}])
prompt = provider.calls[0]["messages"][0]["content"]
assert "USER: earlier thing" in prompt
assert "and now?" in prompt
@pytest.mark.parametrize("bad", ["", "no json here", "{", "}{"])
async def test_no_shape_of_answer_makes_a_step_raise(tmp_path: Path, bad: str) -> None:
p = _personalizer(tmp_path, bad, bad)
assert await p.classify("x") == {"needs_clarification": False, "domain": ""}
assert await p.extract_and_store_preference("x", "q", "a") is False