@classmethod

    def by_account(cls, name: data.Account) -> 'AccrualAccount':

        for account in cls:

            if name.is_under(account.value.name):

                return account

        raise ValueError(f"unrecognized account {name!r}")

    @classmethod

    def classify(cls, related: core.RelatedPostings) -> 'AccrualAccount':

        for account in cls:

            account_name = account.value.name

            if all(post.account.is_under(account_name) for post in related):

                return account

        raise ValueError("unrecognized account set in related postings")

    @property

    def normalize_amount(self) -> Callable[[T], T]:

        return core.normalize_amount_func(self.value.name)

class AccrualPostings(core.RelatedPostings):

    __slots__ = (

        'accrual_type',

        'end_balance',

        'account',

        'entity',

        'invoice',

    )

    INCONSISTENT = Sentinel()

    def __init__(self,

                 source: Iterable[data.Posting]=(),

                 *,

                 _can_own: bool=False,

    ) -> None:

        super().__init__(source, _can_own=_can_own)

        # The following type declarations tell mypy about values set in the for

        # loop that are important enough to be referenced directly elsewhere.

        self.account = self._single_item(post.account for post in self)

        if isinstance(self.account, Sentinel):

            self.accrual_type: Optional[AccrualAccount] = None

            norm_func: Callable[[T], T] = lambda x: x

            entity_pred: Callable[[data.Posting], bool] = bool

        else:

            self.accrual_type = AccrualAccount.by_account(self.account)

            norm_func = self.accrual_type.normalize_amount

            entity_pred = lambda post: norm_func(post.units).number > 0

        self.entity = self._single_item(self.entities(entity_pred))

        self.end_balance = norm_func(self.balance_at_cost())

    def _single_item(self, seq: Iterable[T]) -> Union[T, Sentinel]:

        items = iter(seq)

        try:

            item1 = next(items)

        except StopIteration:

            all_same = False

        else:

            all_same = all(item == item1 for item in items)

        return item1 if all_same else self.INCONSISTENT

    def entities(self, pred: Callable[[data.Posting], bool]=bool) -> Iterator[MetaValue]:

        return filters.iter_unique(

            post.meta['entity']

            for post in self

            if pred(post) and 'entity' in post.meta

        )

    def make_consistent(self) -> Iterator[Tuple[MetaValue, 'AccrualPostings']]:

        account_ok = isinstance(self.account, str)

        entity_ok = isinstance(self.entity, str)

        # `'/' in self.invoice` is just our heuristic to ensure that the

        # invoice metadata is "unique enough," and not just a placeholder

        # value like "FIXME". It can be refined if needed.

        invoice_ok = isinstance(self.invoice, str) and '/' in self.invoice

        if account_ok and entity_ok and invoice_ok:

            yield (self.invoice, self)

            return

        groups = collections.defaultdict(list)

        for post in self:

            post_invoice = self.invoice if invoice_ok else (

                post.meta.get('invoice') or 'BlankInvoice'

            )

            post_entity = self.entity if entity_ok else (

                post.meta.get('entity') or 'BlankEntity'

            )

            groups[f'{post.account} {post_invoice} {post_entity}'].append(post)

        type_self = type(self)

        for group_key, posts in groups.items():

            yield group_key, type_self(posts, _can_own=True)

    def is_paid(self, default: Optional[bool]=None) -> Optional[bool]:

        if self.accrual_type is None:

            return default

        else:

            return self.end_balance.le_zero()

    def is_zero(self, default: Optional[bool]=None) -> Optional[bool]:

        if self.accrual_type is None:

            return default

        else:

            return self.end_balance.is_zero()

    def since_last_nonzero(self) -> 'AccrualPostings':

        for index, (post, balance) in enumerate(self.iter_with_balance()):

            if balance.is_zero():

                start_index = index

        try:

            empty = start_index == index

        except NameError:

            empty = True

        return self if empty else self[start_index + 1:]

class BaseReport:

    def __init__(self, out_file: TextIO) -> None:

                # Cheap optimization: don't slice and dice groups we're not

                # going to report anyway.

                continue

            elif group.accrual_type is None:

                group = group.since_last_nonzero()

            else:

                # Filter out new accruals after the report date.

                # e.g., cover the case that the same invoices has multiple

                # postings over time, and we don't want to report too-recent

                # ones.

                cutoff_date = self.ods.date - datetime.timedelta(

                    days=group.accrual_type.value.aging_thresholds[-1],

                )

                group = AccrualPostings(

                    post for post in group.since_last_nonzero()

                    if post.meta.date <= cutoff_date

                    or group.accrual_type.normalize_amount(post.units.number) < 0

                )

            if group and not group.is_zero():

                rows.append(group)

        rows.sort(key=lambda related: (

            related.account,

            related[0].meta.date,

            ('\0'.join(related.entities())

             if related.entity is related.INCONSISTENT

             else related.entity),

        ))

        self.ods.write(rows)

        self.ods.save_file(self.out_bin)

class BalanceReport(BaseReport):

    def _report(self, posts: AccrualPostings, index: int) -> Iterable[str]:

        posts = posts.since_last_nonzero()

        date_s = posts[0].meta.date.strftime('%Y-%m-%d')

        if index:

            yield ""

        yield f"{posts.invoice}:"

        yield f"  {posts.balance_at_cost()} outstanding since {date_s}"

class OutgoingReport(BaseReport):

    def __init__(self, rt_client: rt.Rt, out_file: TextIO) -> None:

        super().__init__(out_file)

        self.rt_client = rt_client

        self.rt_wrapper = rtutil.RT(rt_client)

    def _primary_rt_id(self, posts: AccrualPostings) -> rtutil.TicketAttachmentIds:

        rt_ids_count = len(rt_ids)

        if rt_ids_count != 1:

            raise ValueError(f"{rt_ids_count} rt-id links found")

        parsed = rtutil.RT.parse(rt_ids.pop())

        if parsed is None:

            raise ValueError("rt-id is not a valid RT reference")

        else:

            return parsed

    def _report(self, posts: AccrualPostings, index: int) -> Iterable[str]:

        posts = posts.since_last_nonzero()

        try:

            ticket_id, _ = self._primary_rt_id(posts)

            ticket = self.rt_client.get_ticket(ticket_id)

            # Note we only use this when ticket is None.

            errmsg = f"ticket {ticket_id} not found"

        except (ValueError, rt.RtError) as error:

            ticket = None

            errmsg = error.args[0]

        if ticket is None:

            self.logger.error(

                "can't generate outgoings report for %s because no RT ticket available: %s",

                posts.invoice, errmsg,

            )

            return

        try:

            rt_requestor = self.rt_client.get_user(ticket['Requestors'][0])

        except (IndexError, rt.RtError):

            rt_requestor = None

        if rt_requestor is None:

            requestor = ''

            requestor_name = ''

        else:

            requestor_name = (

                rt_requestor.get('RealName')

                or ticket.get('CF.{payment-to}')

                or ''

            )

            requestor = f'{requestor_name} <{rt_requestor["EmailAddress"]}>'.strip()

        balance_s = posts.end_balance.format(None)

        raw_balance = -posts.balance()

        if raw_balance != posts.end_balance:

            balance_s = f'{raw_balance} ({balance_s})'

        contract_links = list(posts.all_meta_links('contract'))

        if contract_links:

    def group_by_meta(cls: Type[RelatedType],

                      postings: Iterable[data.Posting],

                      key: MetaKey,

                      default: Optional[MetaValue]=None,

    ) -> Iterator[Tuple[Optional[MetaValue], RelatedType]]:

        """Relate postings by metadata value

        This method takes an iterable of postings and returns a mapping.

        The keys of the mapping are the values of post.meta.get(key, default).

        The values are RelatedPostings instances that contain all the postings

        that had that same metadata value.

        """

        mapping: DefaultDict[Optional[MetaValue], List[data.Posting]] = collections.defaultdict(list)

        for post in postings:

            mapping[post.meta.get(key, default)].append(post)

        for value, posts in mapping.items():

            yield value, cls(posts, _can_own=True)

    def __repr__(self) -> str:

        return f'<{type(self).__name__} {self._postings!r}>'

    @overload

    def __getitem__(self: RelatedType, index: int) -> data.Posting: ...

    @overload

    def __getitem__(self: RelatedType, s: slice) -> RelatedType: ...

    def __getitem__(self: RelatedType,

                    index: Union[int, slice],

    ) -> Union[data.Posting, RelatedType]:

        if isinstance(index, slice):

            return type(self)(self._postings[index], _can_own=True)

        else:

            return self._postings[index]

    def __len__(self) -> int:

        return len(self._postings)

    def _all_meta_links(self, key: MetaKey) -> Iterator[str]:

        for post in self:

            try:

                yield from post.meta.get_links(key)

            except TypeError:

                pass

    def all_meta_links(self, key: MetaKey) -> Iterator[str]:

        return filters.iter_unique(self._all_meta_links(key))

    def iter_with_balance(self) -> Iterator[Tuple[data.Posting, Balance]]:

        balance = MutableBalance()

        for post in self:

            balance += post.units

            yield post, balance

    def balance(self) -> Balance:

        for _, balance in self.iter_with_balance():

            pass

        try:

            return balance

        except NameError:

            return Balance()

    def balance_at_cost(self) -> Balance:

        balance = MutableBalance()

        for post in self:

            if post.cost is None:

                balance += post.units

            else:

                number = post.units.number * post.cost.number

                balance += data.Amount(number, post.cost.currency)

        return balance

    def meta_values(self,

                    key: MetaKey,

                    default: Optional[MetaValue]=None,

    ) -> Set[Optional[MetaValue]]:

        return {post.meta.get(key, default) for post in self}

class BaseSpreadsheet(Generic[RT, ST], metaclass=abc.ABCMeta):

    """Abstract base class to help write spreadsheets

    This class provides the very core logic to write an arbitrary set of data

    rows to arbitrary output. It calls hooks when it starts writing the

    spreadsheet, starts a new "section" of rows, ends a section, and ends the

    spreadsheet.

    RT is the type of the input data rows. ST is the type of the section

    identifier that you create from each row. If you don't want to use the

    section logic at all, set ST to None and define section_key to return None.

    """

    @abc.abstractmethod

    def section_key(self, row: RT) -> ST:

        """Return the section a row belongs to

    related = core.RelatedPostings([

        testutil.Posting('Income:Donations', -1, key='1'),

        testutil.Posting('Income:Donations', -2, key='2'),

    ])

    assert related.meta_values('key', '') == {'1', '2'}

def test_meta_values_many_types():

    expected = {

        datetime.date(2020, 4, 1),

        Decimal(42),

        testutil.Amount(5),

        'rt:42',

    }

    related = core.RelatedPostings(

        testutil.Posting('Income:Donations', -index, key=value)

        for index, value in enumerate(expected)

    )

    assert related.meta_values('key') == expected

@pytest.mark.parametrize('count', range(3))

def test_all_meta_links_zero(count):

    related = core.RelatedPostings(testutil.Posting(

        'Income:Donations', -n, testkey=str(n), _meta_type=data.Metadata,

    ) for n in range(count))

    assert next(related.all_meta_links('approval'), None) is None

def test_all_meta_links_singletons():

    related = core.RelatedPostings(testutil.Posting(

        'Income:Donations', -10, statement=value, _meta_type=data.Metadata,

    ) for value in itertools.chain(

        testutil.NON_LINK_METADATA_STRINGS,

        testutil.LINK_METADATA_STRINGS,

        testutil.NON_STRING_METADATA_VALUES,

    ))

    assert set(related.all_meta_links('statement')) == testutil.LINK_METADATA_STRINGS

def test_all_meta_links_multiples():

    related = core.RelatedPostings(testutil.Posting(

        'Income:Donations', -10, approval=' '.join(value), _meta_type=data.Metadata,

    ) for value in itertools.permutations(testutil.LINK_METADATA_STRINGS, 2))

    assert set(related.all_meta_links('approval')) == testutil.LINK_METADATA_STRINGS

def test_all_meta_links_preserves_order():

    related = core.RelatedPostings(testutil.Posting(

        'Income:Donations', -10, approval=c, _meta_type=data.Metadata,

    ) for c in '121323')

    assert list(related.all_meta_links('approval')) == list('123')

def test_group_by_meta_zero():

    assert not list(core.RelatedPostings.group_by_meta([], 'metacurrency'))

def test_group_by_meta_one(credit_card_cycle):

    posting = next(post for post in data.Posting.from_entries(credit_card_cycle)

                   if post.account.is_credit_card())

    actual = core.RelatedPostings.group_by_meta([posting], 'metacurrency')

    assert set(key for key, _ in actual) == {'USD'}

def test_group_by_meta_many(two_accruals_three_payments):

    postings = [post for post in data.Posting.from_entries(two_accruals_three_payments)

                if post.account == 'Assets:Receivable:Accounts']

    actual = dict(core.RelatedPostings.group_by_meta(postings, 'metacurrency'))

    assert set(actual) == {'USD', 'EUR'}

    for key, group in actual.items():

        assert 2 <= len(group) <= 3

        assert group.balance().is_zero()

def test_group_by_meta_many_single_posts(two_accruals_three_payments):

    postings = [post for post in data.Posting.from_entries(two_accruals_three_payments)

                if post.account == 'Assets:Receivable:Accounts']

    actual = dict(core.RelatedPostings.group_by_meta(postings, 'metanumber'))

    assert set(actual) == {post.units.number for post in postings}

    assert len(actual) == len(postings)