Model/tests/orchestration/test_ingestion_prediction.py

"""IngestionOrchestrator predicts an EPC for an EPC-less Property and persists it
to the predicted slot (ADR-0031 slice-5e). Tested against fakes — no IO. The
prediction collaborators are optional; when unwired, ingestion is unchanged."""

from __future__ import annotations

import json
from datetime import date
from pathlib import Path
from typing import Any, Optional

from datatypes.epc.domain.epc_property_data import EpcPropertyData
from datatypes.epc.domain.mapper import EpcPropertyDataMapper
from domain.epc_prediction.comparable_properties import Comparable
from domain.epc_prediction.epc_prediction import EpcPrediction
from domain.epc_prediction.prediction_target import PredictionTargetAttributes
from domain.geospatial.coordinates import Coordinates
from domain.geospatial.planning_restrictions import PlanningRestrictions
from domain.geospatial.spatial_reference import SpatialReference
from domain.property.property import Property, PropertyIdentity
from orchestration.ingestion_orchestrator import IngestionOrchestrator
from repositories.geospatial.geospatial_repository import GeospatialRepository
from tests.orchestration.fakes import (
    FakeEpcRepo,
    FakePropertyRepo,
    FakeSolarRepo,
    FakeUnitOfWork,
)

_JSON_SAMPLES = Path(__file__).resolve().parents[2] / "backend/epc_api/json_samples"


def _epc() -> EpcPropertyData:
    raw: dict[str, Any] = json.loads(
        (_JSON_SAMPLES / "RdSAP-Schema-21.0.0" / "epc.json").read_text()
    )
    return EpcPropertyDataMapper.from_api_response(raw)


def _property(uprn: Optional[int], postcode: str = "A0 0AA") -> Property:
    return Property(
        identity=PropertyIdentity(
            portfolio_id=1, postcode=postcode, address="1 Some Street", uprn=uprn
        )
    )


class _FakeEpcFetcher:
    def __init__(self, epc: Optional[EpcPropertyData]) -> None:
        self.epc = epc

    def get_by_uprn(self, uprn: int) -> Optional[EpcPropertyData]:
        return self.epc


class _FakeGeospatialRepo(GeospatialRepository):
    def __init__(self, coordinates: Optional[Coordinates]) -> None:
        self._reference = (
            SpatialReference(coordinates=coordinates, restrictions=PlanningRestrictions())
            if coordinates is not None
            else None
        )

    def coordinates_for(self, uprn: int) -> Optional[Coordinates]:
        return self._reference.coordinates if self._reference is not None else None

    def spatial_for(self, uprn: int) -> Optional[SpatialReference]:
        return self._reference


class _FakeSolarFetcher:
    def get_building_insights(
        self, longitude: float, latitude: float
    ) -> dict[str, Any]:
        return {}


class _FakeComparablesRepo:
    def __init__(self, candidates: list[Comparable]) -> None:
        self._candidates = candidates
        self.searched: list[str] = []

    def candidates_for(self, postcode: str) -> list[Comparable]:
        self.searched.append(postcode)
        return self._candidates


class _FakeAttributesReader:
    def __init__(self, attributes: PredictionTargetAttributes) -> None:
        self._attributes = attributes

    def attributes_for(self, property_id: int) -> PredictionTargetAttributes:
        return self._attributes


def _cohort() -> list[Comparable]:
    # Three same-type neighbours (property_type "0"), distinct addresses so the
    # dedupe keeps all three.
    return [
        Comparable(
            epc=_epc(),
            certificate_number=f"CERT-{i}",
            address=f"{i} Some Street",
            registration_date=date(2023, 1, i + 1),
        )
        for i in range(3)
    ]


def test_epc_less_property_is_predicted_and_persisted_to_the_predicted_slot() -> None:
    # Arrange — no lodged EPC, a known property type, and a same-type cohort.
    epc_repo = FakeEpcRepo()
    uow = FakeUnitOfWork(
        property=FakePropertyRepo({10: _property(uprn=12345)}),
        epc=epc_repo,
        solar=FakeSolarRepo(),
    )
    comparables_repo = _FakeComparablesRepo(_cohort())
    orchestrator = IngestionOrchestrator(
        unit_of_work=lambda: uow,
        epc_fetcher=_FakeEpcFetcher(None),
        geospatial_repo=_FakeGeospatialRepo(
            Coordinates(longitude=-0.1, latitude=51.5)
        ),
        solar_fetcher=_FakeSolarFetcher(),
        comparables_repo=comparables_repo,
        prediction_attributes_reader=_FakeAttributesReader(
            PredictionTargetAttributes(property_type="0")
        ),
        epc_prediction=EpcPrediction(),
    )

    # Act
    orchestrator.run([10])

    # Assert — a prediction was synthesised from the postcode cohort and persisted
    # to the predicted slot; the lodged slot stays empty.
    assert comparables_repo.searched == ["A0 0AA"]
    predicted = epc_repo.get_predicted_for_property(10)
    assert predicted is not None
    assert predicted.property_type == "0"
    assert epc_repo.get_for_property(10) is None
    assert uow.commits == 1


def test_unknown_property_type_gates_the_property_out_of_prediction() -> None:
    # Arrange — EPC-less, but the property type could not be resolved.
    epc_repo = FakeEpcRepo()
    uow = FakeUnitOfWork(
        property=FakePropertyRepo({10: _property(uprn=12345)}),
        epc=epc_repo,
        solar=FakeSolarRepo(),
    )
    comparables_repo = _FakeComparablesRepo(_cohort())
    orchestrator = IngestionOrchestrator(
        unit_of_work=lambda: uow,
        epc_fetcher=_FakeEpcFetcher(None),
        geospatial_repo=_FakeGeospatialRepo(Coordinates(longitude=-0.1, latitude=51.5)),
        solar_fetcher=_FakeSolarFetcher(),
        comparables_repo=comparables_repo,
        prediction_attributes_reader=_FakeAttributesReader(
            PredictionTargetAttributes(property_type=None)
        ),
        epc_prediction=EpcPrediction(),
    )

    # Act
    orchestrator.run([10])

    # Assert — gated out: no cohort fetched, nothing predicted or persisted.
    assert comparables_repo.searched == []
    assert epc_repo.get_predicted_for_property(10) is None
    assert epc_repo.get_for_property(10) is None


def test_a_lodged_epc_is_not_predicted_over() -> None:
    # Arrange — a real EPC is fetched, so prediction must not run.
    lodged = _epc()
    epc_repo = FakeEpcRepo()
    uow = FakeUnitOfWork(
        property=FakePropertyRepo({10: _property(uprn=12345)}),
        epc=epc_repo,
        solar=FakeSolarRepo(),
    )
    comparables_repo = _FakeComparablesRepo(_cohort())
    orchestrator = IngestionOrchestrator(
        unit_of_work=lambda: uow,
        epc_fetcher=_FakeEpcFetcher(lodged),
        geospatial_repo=_FakeGeospatialRepo(Coordinates(longitude=-0.1, latitude=51.5)),
        solar_fetcher=_FakeSolarFetcher(),
        comparables_repo=comparables_repo,
        prediction_attributes_reader=_FakeAttributesReader(
            PredictionTargetAttributes(property_type="0")
        ),
        epc_prediction=EpcPrediction(),
    )

    # Act
    orchestrator.run([10])

    # Assert — the lodged EPC is saved; no cohort fetch, no predicted slot.
    assert comparables_repo.searched == []
    assert epc_repo.get_for_property(10) == lodged
    assert epc_repo.get_predicted_for_property(10) is None