Raven-1/tests/test_simulation_reference.py
tzteyang ad2f054dd4
feat(*): add an experimental Curator that cultivates harness-of-harnesses workers (#802)
## Summary

Adds `experimental/`, a first version of worker RSI: a Curator that
shapes a
Raven worker's harness from the task, the materials handed to it and
feedback,
the way a new employee is trained.

The worker is a harness-of-harnesses. A root Raven harness faces the
user and
calls child harnesses, each a full Raven agent hosted over ACP, through
the
nodes of its playbooks. Every harness has the same four strategy facets
(memory, planning, capability, action). The Curator reads the actual
harnesses
and generates code and config on those facets at both levels: the root
candidate sets the playbooks and writes per-node requirements, each
child is
curated against the requirements of its nodes, and the set is checked on
copies and activated together.

Layout:
- `curator/`: staged generation (understand, select, design, implement,
repair), two-level composition, harness contracts, and `raven_adapter/`,
  the only layer that touches a worker's Raven runtime
- `analyst/`: turns an evaluator's words and execution records into
  observable behavior requirements
- `iteration/`: the generic loop (onboarding curation, then trial,
review and
  curation each round) and a CLI for a human evaluator
- `simulation/`: a simulated travel agency that runs the loop, plus one
  showcase case (`simulation/cases/s0925c/`: case notes, the exported
  transcript split by phase, and the round 1 and round 3 decks)
- `docs/design.md` (structure, curation flow, composition sequence, the
Raven
  connection) and `docs/rsi-iteration.md` (the loop and the Analyst)

Key decisions:
- No change under `raven/`, and `raven/` does not import
`experimental/`.
Generated artifacts land on Raven's existing extension points. The few
Raven
internals the adapter reads are listed in `docs/design.md` section 5.4.
- The Curator explores a read-only source snapshot; the caller names the
evaluation side to withhold from it, so the generic layers name no
scenario.
- The single-layer owner analysis is kept only to reproduce the showcase
case
  and is marked for removal; the two-layer Analyst is the default.
- The two decks are `.pptx` files of about 230 KB each. They pass
  `check-large-files`; flagging them in case they count as report assets
under AGENTS.md section 7, in which case they can move out of git behind
a
  link.
- One follow-up for Raven 0.2.2: playbooks in prompt and rounds modes
ship no
  fixed nodes, so node requirements are read from dag playbooks only.

## Type

- [ ] Fix
- [x] Feature
- [ ] Docs
- [ ] CI / tooling
- [ ] Refactor
- [ ] Other

## Verification

On the branch rebased onto the latest `main`:

- `uv run pytest tests/test_harness_curator_*.py tests/test_analyst_*.py
tests/test_iteration_*.py tests/test_simulation_*.py -q`: 368 passed
- `uv run pytest -m integration
tests/integration/test_harness_curator_*.py --deselect
tests/integration/test_harness_curator_real_llm.py -q`
  (offline, stub providers): 57 passed
- `uv run pytest -q` (whole suite, with the host proxy variables unset
and
`PYTHONTZPATH` set): 26692 passed, 2 failed. One was an order-dependent
test
  of this PR and is fixed here; the other,

`tests/test_subagent_node_runtime.py::test_what_cannot_be_read_names_nothing`,
  fails only because the machine runs as root and is unrelated.
- `uv run ruff check` and `uv run ruff format --check` on the touched
files:
  clean
- `scripts/check_large_files.py origin/main..HEAD`,
  `scripts/check_source_language.py origin/main..HEAD`,
  `scripts/check_commit_messages.py origin/main..HEAD` and
  `npx commitlint --from origin/main --to HEAD`: all exit 0

- [x] Relevant tests pass locally
- [x] Relevant lint / type checks pass locally
- [x] User-facing docs or screenshots are updated when needed

## Risk

No user-visible change for Raven: `experimental/` is not imported by
`raven/` and is not packaged in the wheel (`packages = ["raven"]`).
`CONTEXT.md` gains one glossary entry. The Curator's exploration offers
a
shell only under an OS sandbox executor. Rollback is a revert of the
squash
commit; nothing else depends on `experimental/`.

- [x] Security impact considered
- [x] Backward compatibility considered
- [x] Rollback path is clear for risky changes

## Related Issues

N/A

---------

Co-authored-by: Yuqi Yang <yiqiyang524@gmail.com>
Co-authored-by: Claude (claude-opus-5-5) <noreply@anthropic.com>
2026-09-26 21:00:58 +08:00

101 lines
5.2 KiB
Python

"""Figures and deck facts computed for the owner from the scenario's own materials and the played card."""
import shutil
from pptx import Presentation
from experimental.curator.raven_adapter.worker import Execution
from experimental.iteration.protocols import Exchange
from experimental.simulation.cards import Trip
from experimental.simulation.reference import Rules, deck, prices, quoted, references
from experimental.simulation.scenario import BUNDLED, Scenario
TRAVEL = BUNDLED / "travel_agency"
def said(user, reply, files=()):
rows = [{"kind": "runner.event", "event_type": "Text", "event": {"content": reply}}]
return Exchange(user, Execution("t1", [], rows, {}, "a", tuple(str(path) for path in files)))
def trip(start, nights, adults, children=(), budget=99_000):
end = (start[0], start[1] + nights)
return Trip(
"guest", "origin", "somewhere", start, end, nights, adults, tuple(children), (), budget, "139 0571 6628"
)
def loaded():
scenario = Scenario.load(TRAVEL)
return scenario, Rules.load(scenario)
def test_rules_read_prices_seasons_policy_deck_name_and_template_from_the_materials():
_, rules = loaded()
value, comfort, premium = rules.products.values()
assert [product.base for product in (value, comfort, premium)] == [3, 4, 5]
assert [product.extra for product in (value, comfort, premium)] == [260, 800, 2600]
assert [value.child, comfort.child, premium.child] == [True, True, False]
assert comfort.bands == ((1, 1), (2, 3), (4, 8)) and premium.bands == ((1, 1), (2, 2), (3, 6))
assert [prices[1] for prices in comfort.prices.values()] == [3500, 3900, 4600]
off, shoulder, peak = (rules.season(day) for day in ((12, 1), (10, 8), (10, 1)))
assert len({off, shoulder, peak}) == 3
assert rules.season((2, 28)) == off and rules.season((11, 15)) == shoulder and rules.season((10, 7)) == peak
assert (rules.insurance, rules.child_age, rules.child_ratio) == (60, 12, 70)
assert rules.fonts == {"Noto Sans CJK SC", "Noto Serif CJK SC"} and rules.cover != rules.contact
assert rules.quote.fullmatch("HL-Q-1027-4P") and not rules.quote.fullmatch("HL-Q-127-4P")
def test_the_card_gets_every_products_price_as_the_price_list_and_policy_give_it():
scenario, rules = loaded()
sop = trip((12, 2), 4, 3)
comfort = list(rules.products)[1]
assert prices(rules, sop)["by_product"][comfort]["party_total"] == 10_500
assert "10,500" in scenario.text("service-sop")
family = prices(rules, trip((10, 22), 4, 2, (9, 14), budget=7000))
value_row, _, premium_row = family["by_product"].values()
assert (value_row["per_person"], value_row["child_price"], value_row["extra_nights_total"]) == (800, 560, 1040)
assert value_row["children_at_child_price"] == 1 and value_row["party_total"] == 4000
assert premium_row["nights_short_of_base"] == 1 and "child_price" not in premium_row
assert (family["budget_per_person_per_night"], family["insurance_total"]) == (438, 240)
assert family["quote_number"] == "HL-Q-1022-4P"
crowd = prices(rules, trip((10, 22), 4, 9))["by_product"]
assert all(isinstance(row, str) for row in crowd.values())
def test_the_agencys_sample_deck_reads_clean_and_the_bare_template_shows_its_placeholders(tmp_path):
scenario, rules = loaded()
number = "HL-Q-0512-2P"
sample = tmp_path / f"{rules.deck_name}-{number}.pptx"
shutil.copy(scenario.materials["plan-deck-sample"] / "sample.pptx", sample)
exchanges = [said("Quote", f"Quote {number}"), said("Deck?", "Here it is.", [sample])]
facts = deck(rules, exchanges)
assert facts["file"] in facts["names_per_spec"] and facts["opens"] and facts["aspect"] == 1.778
assert facts["fonts_outside_template"] == [] and facts["slides_with_motion"] == 0
assert (facts["first_slide_layout"], facts["last_slide_layout"]) == (rules.cover, rules.contact)
assert facts["template_placeholders_left"] == [] and facts["quote_numbers_in_deck"] == [number]
template = tmp_path / "t" / sample.name
template.parent.mkdir()
shutil.copy(scenario.materials["plan-deck-template"] / "template.pptx", template)
exchanges[-1] = said("Deck?", "Here it is.", [template])
assert deck(rules, exchanges)["template_placeholders_left"] == list(rules.words["placeholders"])
blank = Presentation()
blank.slides.add_slide(blank.slide_layouts[0]).shapes.title.text = number
fresh = tmp_path / "fresh.pptx"
blank.save(fresh)
exchanges[-1] = said("Deck?", "Here it is.", [fresh])
facts = deck(rules, exchanges)
assert facts["aspect"] == 1.333 and facts["fonts_outside_template"] and facts["first_slide_layout"] != rules.cover
broken = tmp_path / "broken.pptx"
broken.write_bytes(b"not a deck")
exchanges[-1] = said("Deck?", "Here it is.", [broken])
assert deck(rules, exchanges)["opens"] is False
def test_references_hold_prices_only_for_a_card_with_a_trip_and_deck_facts_only_after_a_deck():
_, rules = loaded()
exchanges = [said("Hi", "Quote HL-Q-0925-4P, then HL-Q-0925-4P again and HL-Q-1022-4P.")]
assert quoted(exchanges, rules.quote) == ["HL-Q-0925-4P", "HL-Q-1022-4P"]
assert set(references(rules, exchanges, trip((10, 22), 4, 4))) == {"prices"}
assert references(rules, exchanges, None) == {}