SFM/tests/unit/test_column_registry.py

from __future__ import annotations

import json
from pathlib import Path

import pytest

from app.pipeline.column_registry import (
    COLUMNS,
    MAPPING_DATA_FILE,
    AmbiguousColumnNameError,
    ColumnRegistry,
    ColumnSchemaError,
)


def test_registry_keeps_stable_file_name_column_id() -> None:
    assert COLUMNS.index_of("legacy_001") == 1
    assert COLUMNS.meta(1).name == "Имя XML файла"


def test_registry_covers_all_report_columns_without_gaps() -> None:
    indexes = tuple(column.index for column in COLUMNS.ordered_columns())

    assert indexes == tuple(range(1, 205))
    assert all(column.block for column in COLUMNS.ordered_columns())
    assert all(column.report_group for column in COLUMNS.ordered_columns())


def test_registry_assigns_card_columns_missing_from_original_plan() -> None:
    assert COLUMNS.block("foreign_card") == tuple(range(97, 107))


def test_registry_explicitly_preserves_structured_lookup_compatibility() -> None:
    assert COLUMNS.meta(180).allow_short_lookup is True
    assert COLUMNS.meta(153).allow_short_lookup is False


def test_registry_rejects_ambiguous_display_name_lookup() -> None:
    assert COLUMNS.indices_named("ОКАТО") == (4, 171)

    with pytest.raises(AmbiguousColumnNameError):
        COLUMNS.index_named("ОКАТО")


def test_registry_report_groups_are_contiguous_and_cover_schema() -> None:
    covered_indexes: list[int] = []

    for group in COLUMNS.report_groups():
        assert group.indexes == tuple(range(group.start_index, group.end_index + 1))
        covered_indexes.extend(group.indexes)

    assert covered_indexes == list(range(1, 205))


def test_registry_splits_money_and_participant_header_groups() -> None:
    groups = [
        (group.group_id, group.start_index, group.end_index, group.title)
        for group in COLUMNS.report_groups()
        if 14 <= group.start_index <= 154
    ]

    assert groups == [
        ("operation_parameters", 14, 34, "Параметры операции"),
        (
            "transfer_settlement",
            35,
            47,
            "Сведения о переводах денежных средств, в том числе электронных денежных средств",
        ),
        ("cash_receipt", 48, 61, "Сведения о месте приема наличных денежных средств"),
        ("transfer_status", 62, 62, "Статус перевода денежных средств"),
        ("cash_payment", 63, 76, "Сведения о месте выдачи наличных денежных средств"),
        ("esp_authorization", 77, 78, "Сведения о месте авторизации ЭСП"),
        (
            "own_account_cash",
            79,
            96,
            "Сведения о внесении наличных денежных средств на свой банковский счет "
            "или о получении наличных денежных средств со своего банковского счета "
            "у одного оператора по переводу денежных средств",
        ),
        (
            "foreign_card_operation",
            97,
            106,
            "Сведения об операции с использованием платежной карты иностранного банка",
        ),
        ("operation_basis", 107, 107, "Основание совершения операции"),
        ("participant_base", 108, 113, "Сведения об участниках операции"),
        ("participant_fio", 114, 118, "ФИО участника операции"),
        (
            "participant_identifiers",
            119,
            127,
            "Идентификационные сведения участника операции",
        ),
        (
            "participant_identity_document",
            128,
            135,
            "Сведения о документе, удостоверяющем личность",
        ),
        (
            "participant_stay_document",
            136,
            140,
            "Сведения о документе, подтверждающем право на пребывание "
            "(проживание) в Российской Федерации",
        ),
        ("participant_address", 141, 151, "Адрес участника операции"),
        (
            "participant_state_registration",
            152,
            153,
            "Место государственной регистрации участника операции",
        ),
        (
            "participant_extra",
            154,
            154,
            "Дополнительная информация об участнике операции",
        ),
    ]


@pytest.mark.parametrize(
    ("field", "invalid_value"),
    (
        ("index", True),
        ("column_id", None),
        ("column_name", ["Имя"]),
        ("source_scope", {"scope": "any"}),
        ("structured_order", False),
    ),
)
def test_registry_rejects_invalid_scalar_types(
    tmp_path: Path, field: str, invalid_value: object
) -> None:
    data = _mapping_data()
    data[0][field] = invalid_value

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def test_registry_rejects_unknown_report_group(tmp_path: Path) -> None:
    data = _mapping_data()
    data[0]["report_group"] = "unknown_group"

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def test_registry_rejects_unknown_business_block(tmp_path: Path) -> None:
    data = _mapping_data()
    data[0]["block"] = "unknown_block"

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def test_registry_rejects_duplicate_column_ids_and_index_gaps(
    tmp_path: Path,
) -> None:
    data = _mapping_data()
    data[1]["column_id"] = data[0]["column_id"]
    data[1]["index"] = 3

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def test_registry_rejects_structured_group_without_aggregate(
    tmp_path: Path,
) -> None:
    data = _mapping_data()
    aggregate = next(item for item in data if item["index"] == 50)
    aggregate["structured_role"] = "component"
    aggregate["structured_order"] = 10
    aggregate["structured_value"] = False

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def test_registry_rejects_duplicate_structured_component_order(
    tmp_path: Path,
) -> None:
    data = _mapping_data()
    first = next(item for item in data if item["index"] == 51)
    second = next(item for item in data if item["index"] == 52)
    second["structured_order"] = first["structured_order"]

    with pytest.raises(ColumnSchemaError):
        ColumnRegistry.load(_write_mapping(tmp_path, data))


def _mapping_data() -> list[dict[str, object]]:
    return json.loads(MAPPING_DATA_FILE.read_text(encoding="utf-8"))


def _write_mapping(tmp_path: Path, data: list[dict[str, object]]) -> Path:
    path = tmp_path / "mapping_table.json"
    path.write_text(json.dumps(data, ensure_ascii=False), encoding="utf-8")
    return path