"""Behaviour of the Optimiser core: a grouped-knapsack MILP over per-Option
role-1 scores (ADR-0016). Picks at most one Option per Recommendation (disjoint
groups, no cross-group constraints) to maximise total SAP gain subject to the
Scenario budget. This is the warm-start *signal* — the truthful figure comes
from the whole-package re-score + repair (a later slice); here we test the
selection with synthetic scores and no calculator.
"""

from __future__ import annotations

from typing import Sequence

from datatypes.epc.domain.epc_property_data import (
    BuildingPartIdentifier,
    EpcPropertyData,
)
from domain.modelling.optimisation.optimiser import (
    MeasureDependency,
    OptimisedPackage,
    ScoredOption,
    optimise,
    optimise_min_cost,
    optimise_package,
)
from domain.modelling.scoring.package_scorer import Score
from domain.modelling.recommendation import Cost, MeasureOption
from domain.modelling.simulation import (
    BuildingPartOverlay,
    EpcSimulation,
    VentilationOverlay,
)
from tests.domain.sap10_calculator.worksheet._elmhurst_worksheet_000490 import (
    build_epc,
)


def _scored(measure_type: str, *, gain: float, cost: float) -> ScoredOption:
    return ScoredOption(
        option=MeasureOption(
            measure_type=measure_type,
            description=measure_type,
            overlay=EpcSimulation(),
            cost=Cost(total=cost, contingency_rate=0.0),
        ),
        sap_gain=gain,
    )


# Distinguishable overlays so the stub scorer can attribute a true gain per
# measure (wall / roof / floor) regardless of the role-1 signal.
_WALL_OVERLAY = EpcSimulation(
    building_parts={
        BuildingPartIdentifier.MAIN: BuildingPartOverlay(wall_insulation_type=2)
    }
)
_ROOF_OVERLAY = EpcSimulation(
    building_parts={
        BuildingPartIdentifier.MAIN: BuildingPartOverlay(roof_insulation_thickness=300)
    }
)
_FLOOR_OVERLAY = EpcSimulation(
    building_parts={
        BuildingPartIdentifier.MAIN: BuildingPartOverlay(floor_insulation_thickness=100)
    }
)


def _scored_overlay(
    measure_type: str, *, gain: float, cost: float, overlay: EpcSimulation
) -> ScoredOption:
    return ScoredOption(
        option=MeasureOption(
            measure_type=measure_type,
            description=measure_type,
            overlay=overlay,
            cost=Cost(total=cost, contingency_rate=0.0),
        ),
        sap_gain=gain,
    )


class _StubScorer:
    """A deterministic stand-in for PackageScorer: the package SAP is a base
    plus a fixed *true* gain per measure present (by overlay field), decoupled
    from the role-1 signal — so the repair loop is exercised without the
    calculator (ADR-0016)."""

    def __init__(self, *, base: float, wall: float, roof: float, floor: float) -> None:
        self._base = base
        self._wall = wall
        self._roof = roof
        self._floor = floor

    def score(
        self, baseline: EpcPropertyData, simulations: Sequence[EpcSimulation]
    ) -> Score:
        sap = self._base
        for sim in simulations:
            part = sim.building_parts[BuildingPartIdentifier.MAIN]
            if part.wall_insulation_type is not None:
                sap += self._wall
            if part.roof_insulation_thickness is not None:
                sap += self._roof
            if part.floor_insulation_thickness is not None:
                sap += self._floor
        return Score(sap_continuous=sap, co2_kg_per_yr=0.0, primary_energy_kwh_per_yr=0.0)


def _selected_types(selection: list[ScoredOption]) -> set[str]:
    return {scored.option.measure_type for scored in selection}


def test_grouped_knapsack_maximises_gain_within_budget() -> None:
    # Arrange — wall group has two mutually-exclusive options; roof + floor one
    # each. EWI has the best gain but is unaffordable alongside the rest.
    groups: list[list[ScoredOption]] = [
        [
            _scored("external_wall_insulation", gain=10.0, cost=8000.0),
            _scored("cavity_wall_insulation", gain=6.0, cost=1000.0),
        ],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
        [_scored("suspended_floor_insulation", gain=3.0, cost=2000.0)],
    ]

    # Act
    selection: list[ScoredOption] = optimise(groups, budget=5000.0)

    # Assert — cavity + loft + floor (cost 4500, gain 13) beats any package
    # containing the 8000 EWI option within the 5000 budget.
    assert _selected_types(selection) == {
        "cavity_wall_insulation",
        "loft_insulation",
        "suspended_floor_insulation",
    }


def test_picks_at_most_one_option_per_group() -> None:
    # Arrange — both wall options are individually affordable.
    groups: list[list[ScoredOption]] = [
        [
            _scored("external_wall_insulation", gain=10.0, cost=2000.0),
            _scored("cavity_wall_insulation", gain=6.0, cost=1000.0),
        ],
    ]

    # Act
    selection: list[ScoredOption] = optimise(groups, budget=10000.0)

    # Assert — never both treatments of the same wall; the higher-gain one wins.
    assert len(selection) == 1
    assert _selected_types(selection) == {"external_wall_insulation"}


def test_no_budget_picks_the_best_option_in_every_group() -> None:
    # Arrange
    groups: list[list[ScoredOption]] = [
        [
            _scored("external_wall_insulation", gain=10.0, cost=8000.0),
            _scored("cavity_wall_insulation", gain=6.0, cost=1000.0),
        ],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
    ]

    # Act — None budget = unconstrained.
    selection: list[ScoredOption] = optimise(groups, budget=None)

    # Assert
    assert _selected_types(selection) == {
        "external_wall_insulation",
        "loft_insulation",
    }


def test_budget_too_small_for_any_option_selects_nothing() -> None:
    # Arrange
    groups: list[list[ScoredOption]] = [
        [_scored("cavity_wall_insulation", gain=6.0, cost=1000.0)],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
    ]

    # Act
    selection: list[ScoredOption] = optimise(groups, budget=500.0)

    # Assert — nothing affordable; selecting none is the optimum.
    assert selection == []


def test_no_groups_selects_nothing() -> None:
    # Act / Assert
    assert optimise([], budget=10000.0) == []


def test_within_budget_partial_selection_prefers_the_higher_gain_option() -> None:
    # Arrange — only one of the two fits the budget; pick the affordable best.
    groups: list[list[ScoredOption]] = [
        [_scored("external_wall_insulation", gain=10.0, cost=8000.0)],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
    ]

    # Act
    selection: list[ScoredOption] = optimise(groups, budget=2000.0)

    # Assert — EWI is unaffordable; loft alone is the best within £2000.
    assert _selected_types(selection) == {"loft_insulation"}


# --- optimise_min_cost: least-cost-to-target selection (ADR-0016 amendment) ---


def test_min_cost_picks_the_cheapest_package_that_reaches_the_target() -> None:
    # Arrange — two packages both clear the target gain; one is cheaper.
    groups: list[list[ScoredOption]] = [
        [
            _scored("loft_insulation", gain=10.0, cost=2000.0),
            _scored("external_wall_insulation", gain=15.0, cost=3000.0),
        ],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=10000.0, target_gain=10.0)

    # Assert — least-cost-to-target takes the +10 @ £2000, NOT the higher-gain
    # +15 @ £3000 (no overshoot, surplus budget unspent).
    assert selection is not None
    assert _selected_types(selection) == {"loft_insulation"}


def test_min_cost_combines_groups_to_reach_the_target_at_least_cost() -> None:
    # Arrange — no single option reaches +10; the cheapest combo that does is
    # cavity (+6, £1000) + loft (+4, £1500) = +10 @ £2500, beating EWI (+10,
    # £8000).
    groups: list[list[ScoredOption]] = [
        [
            _scored("cavity_wall_insulation", gain=6.0, cost=1000.0),
            _scored("external_wall_insulation", gain=10.0, cost=8000.0),
        ],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=10000.0, target_gain=10.0)

    # Assert
    assert selection is not None
    assert _selected_types(selection) == {
        "cavity_wall_insulation",
        "loft_insulation",
    }


def test_min_cost_breaks_cost_ties_toward_the_higher_gain() -> None:
    # Arrange — two equally-priced packages both reach the target; prefer the
    # one with more headroom ("recommend more" on a tie).
    groups: list[list[ScoredOption]] = [
        [
            _scored("cavity_wall_insulation", gain=10.0, cost=2000.0),
            _scored("external_wall_insulation", gain=14.0, cost=2000.0),
        ],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=10000.0, target_gain=10.0)

    # Assert
    assert selection is not None
    assert _selected_types(selection) == {"external_wall_insulation"}


def test_min_cost_returns_none_when_target_unreachable_within_budget() -> None:
    # Arrange — the only target-reaching package costs more than the budget.
    groups: list[list[ScoredOption]] = [
        [_scored("external_wall_insulation", gain=10.0, cost=8000.0)],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=5000.0, target_gain=10.0)

    # Assert — infeasible (caller falls back to max-gain).
    assert selection is None


def test_min_cost_returns_none_when_no_package_reaches_the_target() -> None:
    # Arrange — even everything together falls short of the target gain.
    groups: list[list[ScoredOption]] = [
        [_scored("cavity_wall_insulation", gain=6.0, cost=1000.0)],
        [_scored("loft_insulation", gain=3.0, cost=1500.0)],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=None, target_gain=10.0)

    # Assert
    assert selection is None


def test_min_cost_unbudgeted_picks_cheapest_reaching_target_not_everything() -> None:
    # Arrange — no budget cap, but min-cost still means cheapest-to-target, not
    # "install everything".
    groups: list[list[ScoredOption]] = [
        [_scored("cavity_wall_insulation", gain=10.0, cost=1000.0)],
        [_scored("loft_insulation", gain=4.0, cost=1500.0)],
    ]

    # Act — cavity alone (+10 @ £1000) already reaches the target.
    selection = optimise_min_cost(groups, budget=None, target_gain=10.0)

    # Assert — loft is left off; it would only add cost past the target.
    assert selection is not None
    assert _selected_types(selection) == {"cavity_wall_insulation"}


def test_min_cost_non_positive_target_selects_nothing() -> None:
    # Arrange — a target already met (gain 0 needed) is reached by the empty
    # package at zero cost.
    groups: list[list[ScoredOption]] = [
        [_scored("cavity_wall_insulation", gain=6.0, cost=1000.0)],
    ]

    # Act
    selection = optimise_min_cost(groups, budget=5000.0, target_gain=0.0)

    # Assert — the cheapest target-reaching package is the empty one.
    assert selection == []


def test_repair_adds_an_untreated_group_option_to_close_the_undershoot() -> None:
    # Arrange — role-1 under-counts roof (signal 0 → warm-start skips it), but
    # its true re-scored gain (+4) is what closes the target.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=0.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
        [_scored_overlay("suspended_floor_insulation", gain=8.0, cost=1000.0, overlay=_FLOOR_OVERLAY)],
    ]
    scorer = _StubScorer(base=40.0, wall=5.0, roof=4.0, floor=3.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=5000.0,
        target_sap=50.0,
    )

    # Assert — warm-start took wall+floor (re-score 48 < 50); repair added the
    # roof (true +4) to reach 52, the truthful package total.
    types = {scored.option.measure_type for scored in package.selected}
    assert "loft_insulation" in types
    assert types == {
        "cavity_wall_insulation",
        "suspended_floor_insulation",
        "loft_insulation",
    }
    assert abs(package.score.sap_continuous - 52.0) <= 1e-9


def test_no_target_returns_the_warm_start_package_without_repair() -> None:
    # Arrange
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
    ]
    scorer = _StubScorer(base=40.0, wall=5.0, roof=4.0, floor=3.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=None,
        target_sap=None,
    )

    # Assert — no target → no repair; warm-start package re-scored as the truth.
    assert {s.option.measure_type for s in package.selected} == {
        "cavity_wall_insulation"
    }
    assert abs(package.score.sap_continuous - 45.0) <= 1e-9


def test_repair_stops_when_no_affordable_improving_option_remains() -> None:
    # Arrange — the only untreated-group option costs more than the budget left.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=0.0, cost=5000.0, overlay=_ROOF_OVERLAY)],
    ]
    scorer = _StubScorer(base=40.0, wall=5.0, roof=4.0, floor=3.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=1000.0,
        target_sap=50.0,
    )

    # Assert — wall only (re-score 45 < 50); roof unaffordable, so repair stops
    # at the best achievable package rather than overspending.
    assert {s.option.measure_type for s in package.selected} == {
        "cavity_wall_insulation"
    }
    assert abs(package.score.sap_continuous - 45.0) <= 1e-9


# --- optimise_package: least-cost-to-target objective (ADR-0016 amendment) ---


def test_package_stops_at_the_target_and_does_not_overshoot() -> None:
    # Arrange — wall alone already clears the target; max-gain would add roof +
    # floor too. Least-cost-to-target must stop at the wall.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=5.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
        [_scored_overlay("suspended_floor_insulation", gain=5.0, cost=1000.0, overlay=_FLOOR_OVERLAY)],
    ]
    scorer = _StubScorer(base=60.0, wall=10.0, roof=5.0, floor=5.0)

    # Act — target 69 (gain 9); wall (+10 → 70) reaches it for £1000.
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=10000.0,
        target_sap=69.0,
    )

    # Assert — just the wall; roof + floor (which would reach 80) are left off,
    # surplus budget unspent.
    assert _selected_types(package.selected) == {"cavity_wall_insulation"}
    assert abs(package.score.sap_continuous - 70.0) <= 1e-9


def test_package_falls_back_to_max_gain_when_target_unreachable() -> None:
    # Arrange — even all three measures (+20 → 80) cannot reach the target.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=5.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
        [_scored_overlay("suspended_floor_insulation", gain=5.0, cost=1000.0, overlay=_FLOOR_OVERLAY)],
    ]
    scorer = _StubScorer(base=60.0, wall=10.0, roof=5.0, floor=5.0)

    # Act — target 90 is out of reach; best effort is the most SAP budget buys.
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=10000.0,
        target_sap=90.0,
    )

    # Assert — max-gain: all three, SAP 80 (below target, best effort).
    assert _selected_types(package.selected) == {
        "cavity_wall_insulation",
        "loft_insulation",
        "suspended_floor_insulation",
    }
    assert abs(package.score.sap_continuous - 80.0) <= 1e-9


def test_package_repairs_when_the_signal_overshoots_the_true_score() -> None:
    # Arrange — the wall's role-1 signal (+10) clears the target gain, so the
    # min-cost warm-start picks it alone; but its true gain is only +5, so the
    # package undershoots and repair must top it up.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=0.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
    ]
    scorer = _StubScorer(base=60.0, wall=5.0, roof=4.0, floor=0.0)

    # Act — target 69 (gain 9). Warm-start {wall} (signal 10) → true 65 < 69 →
    # repair adds the roof (+4) → 69.
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=10000.0,
        target_sap=69.0,
    )

    # Assert
    assert _selected_types(package.selected) == {
        "cavity_wall_insulation",
        "loft_insulation",
    }
    assert abs(package.score.sap_continuous - 69.0) <= 1e-9


# --- Measure Dependency injection (ADR-0016) -------------------------------

_VENT_OVERLAY = EpcSimulation(
    ventilation=VentilationOverlay(
        mechanical_ventilation_kind="EXTRACT_OR_PIV_OUTSIDE"
    )
)


class _VentStubScorer:
    """A stub that adds a fixed gain per wall overlay present and a fixed
    (negative) `vent` contribution when a ventilation overlay is present —
    so the Measure Dependency's effect on the truthful package total and the
    repair decision is exercised without the calculator."""

    def __init__(self, *, base: float, wall: float, roof: float, vent: float) -> None:
        self._base = base
        self._wall = wall
        self._roof = roof
        self._vent = vent

    def score(
        self, baseline: EpcPropertyData, simulations: Sequence[EpcSimulation]
    ) -> Score:
        sap = self._base
        for sim in simulations:
            if sim.ventilation is not None:
                sap += self._vent
            for part in sim.building_parts.values():
                if part.wall_insulation_type is not None:
                    sap += self._wall
                if part.roof_insulation_thickness is not None:
                    sap += self._roof
        return Score(sap_continuous=sap, co2_kg_per_yr=0.0, primary_energy_kwh_per_yr=0.0)


def _ventilation_dependency(*, cost: float) -> MeasureDependency:
    """A forced 'fabric requires ventilation' edge for the tests."""
    return MeasureDependency(
        triggers=frozenset({"cavity_wall_insulation", "external_wall_insulation"}),
        required=ScoredOption(
            option=MeasureOption(
                measure_type="mechanical_ventilation",
                description="mechanical_ventilation",
                overlay=_VENT_OVERLAY,
                cost=Cost(total=cost, contingency_rate=0.0),
            ),
            sap_gain=0.0,
        ),
    )


def test_min_cost_warm_start_avoids_a_wall_whose_forced_ventilation_dooms_it() -> None:
    # Arrange — cavity is dirt cheap (£100) and its role-1 signal (+6) alone
    # reaches the target gain, so a ventilation-BLIND min-cost would pick it.
    # But the wall forces in ventilation at a true/­signal −5, which sinks the
    # package below target. A ventilation-AWARE warm-start prices that −5 into
    # the candidate and instead takes the wall-free loft path.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=6.0, cost=100.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=8.0, cost=1500.0, overlay=_ROOF_OVERLAY)],
    ]
    scorer = _VentStubScorer(base=60.0, wall=6.0, roof=8.0, vent=-5.0)
    dependency = MeasureDependency(
        triggers=frozenset({"cavity_wall_insulation"}),
        required=ScoredOption(
            option=MeasureOption(
                measure_type="mechanical_ventilation",
                description="mechanical_ventilation",
                overlay=_VENT_OVERLAY,
                cost=Cost(total=300.0, contingency_rate=0.0),
            ),
            sap_gain=0.0,  # placeholder; optimise_package scores the real signal
        ),
    )

    # Act — target 66 (gain 6 over the 60 baseline).
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=10000.0,
        target_sap=66.0,
        dependencies=[dependency],
    )

    # Assert — the loft path (true 68, £1500), NOT cavity + forced ventilation:
    # cavity's signal (+6) is cancelled by ventilation (−5) to +1 < target.
    assert _selected_types(package.selected) == {"loft_insulation"}
    assert abs(package.score.sap_continuous - 68.0) <= 1e-9


def test_dependency_injected_when_a_trigger_measure_is_selected() -> None:
    # Arrange — the wall is selected, so its ventilation dependency must be
    # injected before the re-score; ventilation never competes in the pool.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
    ]
    scorer = _VentStubScorer(base=40.0, wall=5.0, roof=4.0, vent=-2.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=None,
        target_sap=None,
        dependencies=[_ventilation_dependency(cost=900.0)],
    )

    # Assert — ventilation is in the package and its negative contribution lands
    # in the truthful total: 40 base + 5 wall − 2 ventilation = 43.
    assert _selected_types(package.selected) == {
        "cavity_wall_insulation",
        "mechanical_ventilation",
    }
    assert abs(package.score.sap_continuous - 43.0) <= 1e-9


def test_dependency_not_injected_without_a_trigger_measure() -> None:
    # Arrange — only loft is selected; the wall-triggered ventilation dependency
    # must not fire.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("loft_insulation", gain=4.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
    ]
    scorer = _VentStubScorer(base=40.0, wall=5.0, roof=4.0, vent=-2.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=None,
        target_sap=None,
        dependencies=[_ventilation_dependency(cost=900.0)],
    )

    # Assert — no trigger, no ventilation; 40 base + 4 roof = 44.
    assert _selected_types(package.selected) == {"loft_insulation"}
    assert abs(package.score.sap_continuous - 44.0) <= 1e-9


def test_wall_dropped_when_it_cannot_be_ventilated_within_budget() -> None:
    # Arrange — cavity (£1000) fits the £1000 budget on its own, but its
    # mandatory ventilation (£900) would bust it. We never blow the budget: a
    # wall we can't afford to ventilate is a wall we can't afford, so it is
    # dropped (the budget is a hard envelope, ventilation is not forced over it).
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
    ]
    scorer = _VentStubScorer(base=40.0, wall=5.0, roof=4.0, vent=-2.0)

    # Act — tight budget; ventilation-aware selection prices the £900 in.
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=1000.0,
        target_sap=None,
        dependencies=[_ventilation_dependency(cost=900.0)],
    )

    # Assert — nothing recommended; the budget is respected and the wall is
    # never left un-ventilated.
    assert package.selected == []


def test_injected_ventilation_penalty_drives_extra_repair() -> None:
    # Arrange — wall (+5) injects ventilation (−2): re-score 43 < target 46.
    # Repair adds the roof (true +4) to reach 47, paying for the ventilation
    # penalty out of the budget the dependency's cost has already eaten into.
    groups: list[list[ScoredOption]] = [
        [_scored_overlay("cavity_wall_insulation", gain=10.0, cost=1000.0, overlay=_WALL_OVERLAY)],
        [_scored_overlay("loft_insulation", gain=0.0, cost=1000.0, overlay=_ROOF_OVERLAY)],
    ]
    scorer = _VentStubScorer(base=40.0, wall=5.0, roof=4.0, vent=-2.0)

    # Act
    package: OptimisedPackage = optimise_package(
        groups=groups,
        scorer=scorer,
        baseline_epc=build_epc(),
        budget=5000.0,
        target_sap=46.0,
        dependencies=[_ventilation_dependency(cost=900.0)],
    )

    # Assert — repair pulled the roof in to clear the target net of ventilation:
    # 40 + 5 wall − 2 vent + 4 roof = 47.
    assert _selected_types(package.selected) == {
        "cavity_wall_insulation",
        "loft_insulation",
        "mechanical_ventilation",
    }
    assert abs(package.score.sap_continuous - 47.0) <= 1e-9