url_authority result shape is inconsistent across URL grammars

[fhightower]

In ioc_finder/ioc_grammars.py, four grammars expose a url_authority named result but with two different shapes:

grammar	parse_string(...).url_authority type
url	str
scheme_less_url	str
url_complete	ParseResults (with url_host inside)
scheme_less_url_complete	ParseResults (with url_host inside)

Reproduced on main:

>>> from ioc_finder import ioc_grammars
>>> ioc_grammars.url.parse_string("https://example.com/path").url_authority
'example.com'                                  # str
>>> ioc_grammars.url_complete.parse_string("https://example.com/path").url_authority
ParseResults(['example.com'], {'url_host': ['example.com']})   # ParseResults

Consumers have to special-case the shape, e.g. on the #369 branch tests/test_urls.py::test_parse_url_helper_handles_scheme_ful_and_scheme_less does:

authority = parsed.url_authority
if not isinstance(authority, str):
    authority = authority[0]

That branch is doing the consumer's job to paper over the grammar API. Pre-existing — not introduced by #369 — but #369's new _parse_url 4-grammar cascade broadens the surface that the asymmetry leaks through, which makes the inconsistency more visible to downstream callers.

Suggested fix

Normalize all four grammars to return url_authority as a ParseResults (with url_host and, where applicable, url_userinfo inside) — or as a str everywhere, whichever matches the design intent. Whichever direction is chosen, the named result should mean the same thing across grammars so callers can rely on a single shape.

Refs

• Restrict scheme-less URL grammar to scheme-less URLs only #369 (where the asymmetry becomes more callable)
• tests/test_urls.py::test_parse_url_helper_handles_scheme_ful_and_scheme_less