conservancy_beancount/conservancy_beancount/data.py

"""Enhanced Beancount data structures for Conservancy

The classes in this module are interface-compatible with Beancount's core data
structures, and provide additional business logic that we want to use
throughout Conservancy tools.
"""
# Copyright © 2020  Brett Smith
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.

import collections
import datetime
import decimal
import functools

from beancount.core import account as bc_account
from beancount.core import amount as bc_amount
from beancount.core import convert as bc_convert
from beancount.core import position as bc_position

from typing import (
    cast,
    Callable,
    Hashable,
    Iterable,
    Iterator,
    MutableMapping,
    Optional,
    Sequence,
    TypeVar,
    Union,
)

from .beancount_types import (
    Directive,
    MetaKey,
    MetaValue,
    Posting as BasePosting,
    Transaction,
)

DecimalCompat = Union[decimal.Decimal, int]

LINK_METADATA = frozenset([
    'approval',
    'check',
    'contract',
    'invoice',
    'purchase-order',
    'receipt',
    'rt-id',
    'statement',
])

class Account(str):
    """Account name string

    This is a string that names an account, like Assets:Bank:Checking
    or Income:Donations. This class provides additional methods for common
    account name parsing and queries.
    """
    __slots__ = ()

    SEP = bc_account.sep

    def is_cash_equivalent(self) -> bool:
        return (
            self.is_under('Assets:') is not None
            and self.is_under('Assets:Prepaid', 'Assets:Receivable') is None
        )

    def is_checking(self) -> bool:
        return self.is_cash_equivalent() and ':Check' in self

    def is_credit_card(self) -> bool:
        return self.is_under('Liabilities:CreditCard') is not None

    def is_opening_equity(self) -> bool:
        return self.is_under('Equity:Funds', 'Equity:OpeningBalance') is not None

    def is_under(self, *acct_seq: str) -> Optional[str]:
        """Return a match if this account is "under" a part of the hierarchy

        Pass in any number of account name strings as arguments. If this
        account is under one of those strings in the account hierarchy, the
        first matching string will be returned. Otherwise, None is returned.

        You can use the return value of this method as a boolean if you don't
        care which account string is matched.

        An account is considered to be under itself:

          Account('Expenses:Tax').is_under('Expenses:Tax') # returns 'Expenses:Tax'

        To do a "strictly under" search, end your search strings with colons:

          Account('Expenses:Tax').is_under('Expenses:Tax:') # returns None
          Account('Expenses:Tax').is_under('Expenses:') # returns 'Expenses:'

        This method does check that all the account boundaries match:

          Account('Expenses:Tax').is_under('Exp') # returns None
        """
        for prefix in acct_seq:
            if self.startswith(prefix) and (
                prefix.endswith(self.SEP)
                or self == prefix
                or self[len(prefix)] == self.SEP
            ):
                return prefix
        return None


class Amount(bc_amount.Amount):
    """Beancount amount after processing

    Beancount's native Amount class declares number to be Optional[Decimal],
    because the number is None when Beancount first parses a posting that does
    not have an amount, because the user wants it to be automatically balanced.

    As part of the loading process, Beancount replaces those None numbers
    with the calculated amount, so it will always be a Decimal. This class
    overrides the type declaration accordingly, so the type checker knows
    that our code doesn't have to consider the possibility that number is
    None.
    """
    number: decimal.Decimal

    # beancount.core._Amount is the plain namedtuple.
    # beancore.core.Amount adds instance methods to it.
    # b.c.Amount.__New__ calls `b.c._Amount.__new__`, which confuses type
    # checking. See <https://github.com/python/mypy/issues/1279>.
    # It works fine if you use super(), which is better practice anyway.
    # So we override __new__ just to call _Amount.__new__ this way.
    def __new__(cls, number: decimal.Decimal, currency: str) -> 'Amount':
        return super(bc_amount.Amount, Amount).__new__(cls, number, currency)


class Metadata(MutableMapping[MetaKey, MetaValue]):
    """Transaction or posting metadata

    This class wraps a Beancount metadata dictionary with additional methods
    for common parsing and query tasks.
    """
    __slots__ = ('meta',)

    def __init__(self, source: MutableMapping[MetaKey, MetaValue]) -> None:
        self.meta = source

    def __iter__(self) -> Iterator[MetaKey]:
        return iter(self.meta)

    def __len__(self) -> int:
        return len(self.meta)

    def __getitem__(self, key: MetaKey) -> MetaValue:
        return self.meta[key]

    def __setitem__(self, key: MetaKey, value: MetaValue) -> None:
        self.meta[key] = value

    def __delitem__(self, key: MetaKey) -> None:
        del self.meta[key]

    def get_links(self, key: MetaKey) -> Sequence[str]:
        try:
            value = self.meta[key]
        except KeyError:
            return ()
        if isinstance(value, str):
            return value.split()
        else:
            raise TypeError("{} metadata is a {}, not str".format(
                key, type(value).__name__,
            ))


class PostingMeta(Metadata):
    """Combined access to posting metadata with its parent transaction metadata

    This lets you access posting metadata through a single dict-like object.
    If you try to look up metadata that doesn't exist on the posting, it will
    look for the value in the parent transaction metadata instead.

    You can set and delete metadata as well. Changes only affect the metadata
    of the posting, never the transaction. Changes are propagated to the
    underlying Beancount data structures.

    Functionally, you can think of this as identical to:

      collections.ChainMap(post.meta, txn.meta)

    Under the hood, this class does a little extra work to avoid creating
    posting metadata if it doesn't have to.
    """
    __slots__ = ('txn', 'index', 'post')

    def __init__(self,
                 txn: Transaction,
                 index: int,
                 post: Optional[BasePosting]=None,
    ) -> None:
        if post is None:
            post = txn.postings[index]
        self.txn = txn
        self.index = index
        self.post = post
        if post.meta is None:
            self.meta = self.txn.meta
        else:
            self.meta = collections.ChainMap(post.meta, txn.meta)

    def __getitem__(self, key: MetaKey) -> MetaValue:
        try:
            return super().__getitem__(key)
        except KeyError:
            if key == 'entity' and self.txn.payee is not None:
                return self.txn.payee
            else:
                raise

    def __setitem__(self, key: MetaKey, value: MetaValue) -> None:
        if self.post.meta is None:
            self.post = self.post._replace(meta={key: value})
            self.txn.postings[self.index] = self.post
            # mypy complains that self.post.meta could be None, but we know
            # from two lines up that it's not.
            self.meta = collections.ChainMap(self.post.meta, self.txn.meta)  # type:ignore[arg-type]
        else:
            super().__setitem__(key, value)

    def __delitem__(self, key: MetaKey) -> None:
        if self.post.meta is None:
            raise KeyError(key)
        else:
            super().__delitem__(key)

    # This is arguably cheating a litttle bit, but I'd argue the date of
    # the parent transaction still qualifies as posting metadata, and
    # it's something we want to access so often it's good to have it
    # within easy reach.
    @property
    def date(self) -> datetime.date:
        return self.txn.date


class Posting(BasePosting):
    """Enhanced Posting objects

    This class is a subclass of Beancount's native Posting class where
    specific fields are replaced with enhanced versions:

    * The `account` field is an Account object
    * The `units` field is our Amount object (which simply declares that the
      number is always a Decimal—see that docstring for details)
    * The `meta` field is a PostingMeta object
    """
    __slots__ = ()

    account: Account
    units: Amount
    cost: Optional[bc_position.Cost]
    # mypy correctly complains that our MutableMapping is not compatible
    # with Beancount's meta type declaration of Optional[Dict]. IMO
    # Beancount's type declaration is a smidge too specific: I think its type
    # declaration should also use MutableMapping, because it would be very
    # unusual for code to specifically require a Dict over that.
    # If it did, this declaration would pass without issue.
    meta: PostingMeta  # type:ignore[assignment]

    @classmethod
    def from_beancount(cls,
                       txn: Transaction,
                       index: int,
                       post: Optional[BasePosting]=None,
    ) -> 'Posting':
        if post is None:
            post = txn.postings[index]
        return cls(
            Account(post.account),
            *post[1:5],
            # see rationale above about Posting.meta
            PostingMeta(txn, index, post), # type:ignore[arg-type]
        )

    @classmethod
    def from_txn(cls, txn: Transaction) -> Iterator['Posting']:
        """Yield an enhanced Posting object for every posting in the transaction"""
        for index, post in enumerate(txn.postings):
            yield cls.from_beancount(txn, index, post)

    @classmethod
    def from_entries(cls, entries: Iterable[Directive]) -> Iterator['Posting']:
        """Yield an enhanced Posting object for every posting in these entries"""
        for entry in entries:
            # Because Beancount's own Transaction class isn't type-checkable,
            # we can't statically check this. Might as well rely on duck
            # typing while we're at it: just try to yield postings from
            # everything, and ignore entries that lack a postings attribute.
            try:
                yield from cls.from_txn(entry)  # type:ignore[arg-type]
            except AttributeError:
                pass


_KT = TypeVar('_KT', bound=Hashable)
_VT = TypeVar('_VT')
class _SizedDict(collections.OrderedDict, MutableMapping[_KT, _VT]):
    def __init__(self, maxsize: int=128) -> None:
        self.maxsize = maxsize
        super().__init__()

    def __setitem__(self, key: _KT, value: _VT) -> None:
        super().__setitem__(key, value)
        for _ in range(self.maxsize, len(self)):
            self.popitem(last=False)


def balance_of(txn: Transaction,
               *preds: Callable[[Account], Optional[bool]],
) -> Amount:
    """Return the balance of specified postings in a transaction.

    Given a transaction and a series of account predicates, balance_of
    returns the balance of the amounts of all postings with accounts that
    match any of the predicates.

    balance_of uses the "weight" of each posting, so the return value will
    use the currency of the postings' cost when available.
    """
    match_posts = [post for post in Posting.from_txn(txn)
                   if any(pred(post.account) for pred in preds)]
    number = decimal.Decimal(0)
    if not match_posts:
        currency = ''
    else:
        weights: Sequence[Amount] = [
            bc_convert.get_weight(post) for post in match_posts
        ]
        number = sum((wt.number for wt in weights), number)
        currency = weights[0].currency
    return Amount(number, currency)

_opening_balance_cache: MutableMapping[str, bool] = _SizedDict()
def is_opening_balance_txn(txn: Transaction) -> bool:
    key = '\0'.join(
        f'{post.account}={post.units}' for post in txn.postings
    )
    try:
        return _opening_balance_cache[key]
    except KeyError:
        pass
    opening_equity = balance_of(txn, Account.is_opening_equity)
    if not opening_equity.currency:
        retval = False
    else:
        rest = balance_of(txn, lambda acct: not acct.is_opening_equity())
        if not rest.currency:
            retval = False
        else:
            retval = abs(opening_equity.number + rest.number) < decimal.Decimal('.01')
    _opening_balance_cache[key] = retval
    return retval