[FatmaMahmoudR]

Closes #192.

ordinal takes the suffix of a decimal number from its last digit, so ordinal("1.0") is "1.0th". A float holding a whole number went down a separate path that used its integer value instead:

>>> p.ordinal("1.0")
'1.0th'
>>> p.ordinal(1.0)  # before
'1.0st'
>>> p.ordinal(2.0)  # before
'2.0nd'

Now any number with a decimal point uses its last digit, so ordinal(1.0) == ordinal("1.0") == "1.0th", the same way the existing test has ordinal(1.1) == ordinal("1.1"). A float's ordinal now always matches its string's, including large floats Python writes in scientific notation, such as 1.1e+21.

For the other case in the issue, ordinal("1."), there's no digit after the point to take the suffix from, so it now raises BadNumValueError instead of returning "1.st". If you'd rather keep "1.st", I can put that branch back.

With both special cases gone, the branching in ordinal comes down to one line.