Implementing class structure with a lot of business logic

portfolio_example_1.py

"""
Portfolio implementation
"""


class Portfolio:
    """Used for backtesting strategies and taking actions through time."""

    def __init__(self, **kwargs):
        self.capital_allocation = capital_allocation
        ...
        self.s3_client = boto3.client("s3")

    @property
    def transactions(self):
        """Records all transactions through portfolio's existence period."""
        if self._transactions.empty:
            self.run()
        return self._transactions

    def get_adjusted_close(self, ticker: str, trading_day: Arrow) -> float:
        """Returns the closing price for a symbol on a given day."""
        # no need to inject the s3_client since we can just depend on self!
        return qm.equities.historical.get_price(self.s3_client, self.s3_bucket, ticker, trading_day)

    def update_positions(self):
        # updates the transactions dataframe, which is port of a portfolio instance
        # that's ok...right?
        return df

    def run(self):
        """
        Iterates through each day of the portfolio time period.

        Add record to positions dataframe and returns series identify if action should be taken if so.
        pull data for equity universe and identify top n stocks
        calculate position size for each stock and number of stocks to purchase
        liquidate all existing positions (add to transactions dataframe)
        enter new positions (add to transactions dataframe)
        Rebalance quarterly, before quarter ending months: end of February, May, August, and November.
        """
        # orchestrates calls to various funcitons and updates "private" variables like _transactions
        pass

"""
Testing our Portfolio implementation
"""


@mock.patch('client.s3_client')
def test_portfolio_update_postions():
    # need to mock our s3_client so that our portfolio can initialize
    portfolio = Portfolio(capital_allocation=100000..., start_date=... end_date=...)
    # set the _transactions property so that our update_positions has the data it needs within
    # this Portfolio instance, hope we remember all the other pieces of state we need to set up
    # to get this test working!
    portfolio._transactions = pd.DataFrame(...)
    portfolio.update_positions()
    expected_positions = ...
    assert expected_positions == portfolio.positions