beancount-cde-importer/beancount_cde_importer/__init__.py

import csv
import re

from datetime import datetime, date
from itertools import islice
from typing import Any, Optional

from beancount.core import data  # type: ignore
from beancount.core.amount import Amount  # type: ignore
from beancount.core.number import Decimal  # type: ignore
from beancount.ingest import cache, importer  # type: ignore


INDEX_DATE = 0
INDEX_TRANSACTION_NUMBER = 1
INDEX_LABEL = 2
INDEX_DEBIT = 3
INDEX_CREDIT = 4
INDEX_DETAIL = 5

END_DATE_REGEX = "Date de fin de téléchargement : ([0-3][0-9]/[0-1][0-9]/[0-9]{4})"
START_DATE_REGEX = "Date de début de téléchargement : ([0-3][0-9]/[0-1][0-9]/[0-9]{4})"


def is_valid_header(header: list[str]) -> bool:
    return (
        header[INDEX_DATE] == "Date"
        and header[INDEX_TRANSACTION_NUMBER] == "Numéro d'opération"
        and header[INDEX_LABEL] == "Libellé"
        and header[INDEX_DEBIT] == "Débit"
        and header[INDEX_CREDIT] == "Crédit"
        and header[INDEX_DETAIL] == "Détail"
    )


def get_date(file: cache._FileMemo, regex: str) -> Optional[date]:
    match: Optional[re.Match] = re.search(regex, file.head())
    if match is None:
        return None
    date_str: Optional[str] = match.group(1)
    if date_str is None:
        return None
    return datetime.strptime(date_str, "%d/%m/%Y").date()


def get_end_date(file: cache._FileMemo) -> Optional[date]:
    return get_date(file, END_DATE_REGEX)


def get_start_date(file: cache._FileMemo) -> Optional[date]:
    return get_date(file, START_DATE_REGEX)


class CDEImporter(importer.ImporterProtocol):
    def __init__(self, account: str):
        self.account = account

    def identify(self, file: cache._FileMemo) -> bool:
        try:
            # NOTE: beancount.ingest.cache._FileMemo handles automatic encoding
            # detection
            lines: list[str] = file.head().splitlines()
            csv_reader = csv.reader(
                lines, delimiter=";", strict=True, quoting=csv.QUOTE_NONE
            )

            # header is actually on the 5th line, the previous ones contain
            # miscellaneous information
            header: Optional[list[str]] = next(islice(csv_reader, 4, None))

            if header is None:
                return False
            return is_valid_header(header)

        except:
            return False

    def file_account(self, file: cache._FileMemo) -> Optional[str]:
        return self.account

    def file_name(self, file: cache._FileMemo) -> Optional[str]:
        return "CaisseEpargne_Statement.csv"

    def file_date(self, file: cache._FileMemo) -> Optional[date]:
        return get_end_date(file)

    def extract(self, file: cache._FileMemo, existing_entries=None) -> list[Any]:
        directives: list[Any] = []

        end_date: Optional[date] = get_end_date(file)
        start_date: Optional[date] = get_start_date(file)
        if end_date is None or start_date is None:
            return directives

        lines: list[str] = file.contents().splitlines()
        csv_reader = csv.reader(
            lines, delimiter=";", strict=True, quoting=csv.QUOTE_NONE
        )

        # first 3 lines are useless
        for _ in range(3):
            next(csv_reader)

        # 4th line is usually the final balance
        row: Optional[list[str]] = next(csv_reader)
        if row is None:
            return directives

        if row[0] == "Solde en fin de période":
            meta = data.new_metadata(file.name, 4)
            balance = Decimal(row[4].replace(",", "."))
            directives.append(
                data.Balance(
                    meta=meta,
                    date=end_date,
                    account=self.account,
                    amount=Amount(balance, "EUR"),
                    tolerance=None,
                    diff_amount=None,
                )
            )

        # skip headings
        next(csv_reader)

        for index, row in enumerate(csv_reader):
            lineno: int = index + 6  # entries start at line 6
            meta = data.new_metadata(file.name, lineno)

            if row[0] == "Solde en début de période":
                balance = Decimal(row[4].replace(",", "."))
                directives.append(
                    data.Balance(
                        meta=meta,
                        date=start_date,
                        account=self.account,
                        amount=Amount(balance, "EUR"),
                        tolerance=None,
                        diff_amount=None,
                    )
                )

                # should be the last line anyway
                continue

            transaction_date: date = datetime.strptime(
                row[INDEX_DATE], "%d/%m/%y"
            ).date()
            label: str = row[INDEX_LABEL]
            debit: str = row[INDEX_DEBIT]
            credit: str = row[INDEX_CREDIT]

            postings: list[data.Posting] = []

            if debit:
                amount = Decimal(debit.replace(",", "."))
                postings.append(
                    data.Posting(
                        self.account, Amount(amount, "EUR"), None, None, None, None
                    )
                )

            if credit:
                amount = Decimal(credit.replace(",", "."))
                postings.append(
                    data.Posting(
                        self.account, Amount(amount, "EUR"), None, None, None, None
                    )
                )

            directives.append(
                data.Transaction(
                    meta,
                    transaction_date,
                    self.FLAG,
                    label,
                    "",
                    data.EMPTY_SET,
                    data.EMPTY_SET,
                    postings,
                )
            )

        return directives