Improve documentation for test of itemized deduction components

[Peter-Metz]

In preparing #2404, @MattHJensen and I found that test_itemded_component_amounts() in test_calculator.py could use some improved documentation. This test compares the calculated amount for a component of itemized deductions (for example c17000) with the amount that results from implementing a haircut for the corresponding component (for example ID_Medical_hc). In theory these values should be the same, but in practice there are small discrepancies in some cases:

    if year == 2017 and cvname == 'c19700':
        atol = 0.016
    elif year == 2017 and cvname == 'c19200':
        atol = 0.010
    elif year == 2017 and cvname == 'c18300':
        atol = 0.009
    else:
        atol = 0.00001

where atol is the discrepancy in millions USD. Moving forward, it would be helpful to document why these margin of errors are necessary and why they might change after data updates.

[jdebacker]

@MattHJensen @Peter-Metz I'm happy to add some inline comments to document this, but I don't know why there tolerances vary for certain variables in certain years. Do you?