model.base: fix id_short validation

In RegEx, the `$` asserts the end of a string or line. When used in
re.match() it asserts the end of the string by default, unless the last
character is a newline symbol (`\n`). Then it asserts the end of the string
*before* the newline. Thus the first regex check doesn't filter valid
id_shorts ending in a newline, which it should.
See https://docs.python.org/3.9/library/re.html#index-2

Using re.fullmatch() instead fixes this issue, making the `^` and `$`
unnecessary as well.

Next to adding a test for this behaviour this commit also simplifies the
check for an alphabetical character at the start of the id_short.
And although both checks could also be merged into one, they are left
separate to be able to report more accurate errors.