Parse month- and day-level consecutives in sets

[lisboarocha]

Fixes #79.

Sets with a range of months or days did not parse:

>>> parse_edtf("[1774-11..1774-12]")
TypeError: Consecutives.__init__() got an unexpected keyword argument 'year'

Two things went wrong in consecutives. The results names inside a month or day bound (year, month, day) reached the token dict, so Consecutives(**kwargs) received them as keyword arguments; and each bound arrived as the list of its parts (['1774', '-', '11']) rather than as a string, which is what year_basic works around for years.

The fix:

• grammar.py: the month and day bounds are wrapped in Combine, so each is one string ("1774-11").
• parser_classes.py: Consecutives.parse_action builds the object from lower and upper only, taking the combined string out of the ParseResults that Combine returns when the bound keeps inner results names.

Tests: four examples added to EXAMPLES in tests/test_parser.py, from the issue and the specification's set forms:

• [1774-11..1774-12] → 1774-11-01 .. 1774-12-31
• [1785-07-30..1785-08-06] → 1785-07-30 .. 1785-08-06
• [1667, 1912-03..1912-05] → 1667-01-01 .. 1912-05-31
• {1912-03..1912-05} → 1912-03-01 .. 1912-05-31

The four failed before the change; with it the suite gives 290 passed, 22 deselected (286 before), and ruff check and ruff format --check pass on the changed files.

[aweakley]

@cameron-simpson please can you look at this one?

[cameron-simpson]

This seems correct to me. Thank you @lisboarocha