Solution: Roman numbers

def roman_to_arabic(roman):
    r2a = {
            "I" : 1,
            "V" : 5,
            "X" : 10,
            "L" : 50,
            "C" : 100,
            "D" : 500,
            "M" : 1000,
    }
    arabic = 0
    for i in range(len(roman)):
        if roman[i] not in r2a:
            raise ValueError(f"Invalid character '{roman[i]}'")
        if i == len(roman) - 1:
            arabic += r2a[roman[i]]
            break

        if roman[i+1] not in r2a:
            raise ValueError(f"Invalid character '{roman[i+1]}'")

        if r2a[roman[i]] < r2a[roman[i+1]]:
            arabic -= r2a[roman[i]]
        else:
            arabic += r2a[roman[i]]

    return arabic

import pytest
from roman_to_arabic import roman_to_arabic

@pytest.mark.parametrize("roman, arabic", [
    ("I", 1),
    ("V", 5),
    ("X", 10),
    ("L", 50),
    ("C", 100),
    ("D", 500),
    ("M", 1000),
    ("II", 2),
    ("III", 3),
    ("IV", 4),
    ("MMMCLXXVIII", 3178),
    ("MMMCMXCIX", 3999),
])
def test_roman_to_arabic(roman, arabic):
    assert roman_to_arabic(roman) == arabic

def test_bad_1():
    with pytest.raises(ValueError) as err:
        roman_to_arabic("Z")
    assert str(err.value) == "Invalid character 'Z'"

def test_bad_2():
    with pytest.raises(ValueError) as err:
        roman_to_arabic("IVZ")
    assert str(err.value) == "Invalid character 'Z'"

#def test_bad_2():
#    with pytest.raises(ValueError) as err:
#        roman_to_arabic("IIII")
#    assert str(err.value) == "Invalid roman number 'IIII'"