Test suite not passing following recent changes

[tepelbaum]

Hi!

Thanks a lot for this great library :). I was using the latest github version but noticed that the test suite is broken. Maybe it could be good to add a github workflow running the test suite? Or at least to fix the test/code to show the correct intent :).

One of the cases I have in mind is (It's not the only one that needs fixing if I am not mistaken)

company_name	company_id	isic	target_type	intensity_metric	scope	emissions_in_scope	reduction_ambition	base_year	start_year	end_year	achieved_reduction	target_status	time_frame	industry	ghg_s1s2	ghg_s3	company_market_cap	investment_value	portfolio_weight	company_cash_equivalents	company_enterprise_value	company_ev_plus_cash	company_total_assets	sbti_validated
Company T	CA0000000020	A12	Absolute	Revenue	S1+S2	100	0.416	2009	2009	2020	0.68	Underway	short	test	28400000	28400000	66682	18472	0.055325267	22005.06	15403.542	62681.08	60680.62	1

One line of the data_test_temperature_score.csv file

Here the end_year is 2020. If we go into

    def get_annual_reduction_rate(self, target: pd.Series) -> Optional[float]:
        """
        Get the annual reduction rate (or None if not available).

        :param target: The target as a row of a dataframe
        :return: The annual reduction
        """

        # 2022-09-01 Bloomberg pointed out need for additional checks in input
        # Here is the original code:
        # if pd.isnull(target[self.c.COLS.REDUCTION_AMBITION]):
        #     return None

        # try:
        #     return target[self.c.COLS.REDUCTION_AMBITION] / float(
        #         target[self.c.COLS.END_YEAR] - target[self.c.COLS.BASE_YEAR]
        #     )
        # except ZeroDivisionError:
        #     raise ValueError(
        #         "Couldn't calculate the annual reduction rate because the start and target year are the "
        #         "same"
        #     )

        # Bloombergs proposal - changed 2022-09-01
        check = pd.isnull(target[self.c.COLS.REDUCTION_AMBITION])
        check = check or pd.isnull(target[self.c.COLS.END_YEAR])
        check = check or pd.isnull(target[self.c.COLS.BASE_YEAR])
        check = check or (target[self.c.COLS.END_YEAR] <= target[self.c.COLS.BASE_YEAR])
        # add check that target is not too old
        check = check or (target[self.c.COLS.END_YEAR] < datetime.datetime.now().year)
        if check:
            return None
        return target[self.c.COLS.REDUCTION_AMBITION] / float(
            target[self.c.COLS.END_YEAR] - target[self.c.COLS.BASE_YEAR]
        )
        # End of BBGs code

We see that the last check puts NaN into self.c.COLS.ANNUAL_REDUCTION_RATE, thus giving a final temperature of 3.2, whereas in the test we find:

  scores = self.temperature_score.calculate(self.data)
        self.assertAlmostEqual(
            scores[
                (scores["company_name"] == "Company T")
                & (scores["scope"] == EScope.S1S2)
            ]["temperature_score"].iloc[0],
            1.77,
            places=2,
            msg="The temp score was incorrect",
        )

1.77 :)

[mountainrambler]

Thanks for bringing this up; and sorry for the very late response.
Obviously, a target with an end date in 2020 would be invalid. So we'll take a look at updating the test file as well as finding some enhanced workflow.