feat: add --ignore-sic to skip misspellings marked with [sic]

[kojiromike]

Closes #3764.

Summary

Adds an opt-in --ignore-sic flag that skips a misspelling immediately followed by the editorial [sic] marker.

• Only the single occurrence preceding the marker is ignored; other misspellings on the same line are still reported.
• A closing quote may sit between the word and the marker, which is the common case when documenting a corrected typo in a changelog, e.g. correct the "wrod" [sic] typo.
• Case-insensitive ([sic]/[SIC]), with optional whitespace before the marker.

Motivation

The existing inline mechanisms are all annotations aimed at the tool: codespell:ignore (whole line), codespell:ignore <word> (by spelling, not position — it can't target a single occurrence and requires naming the typo in a comment), and codespell:ignore-next-line. None of these fit prose where an intentional/quoted misspelling appears, such as a changelog entry announcing a typo fix.

[sic] is a long-standing editorial convention that already belongs in the text on its own merits, so reusing it costs the document nothing and needs no tooling comment. Because it changes default behavior only when explicitly requested, it is gated behind a flag rather than folded into codespell:ignore.

Implementation

• New sic_regex and an --ignore-sic argument (default off).
• The check sits beside the existing backslash-escape false-alarm lookbehind in the match loop, using a positional re.match(line, match.end()) anchor.

Test plan

• New parametrized test_ignore_sic: bare/quoted/whitespace/case variants, the changelog quoted form, single-occurrence scoping, and negative cases ((sic), [sick], intervening word). Asserts off-by-default behavior too.
• ruff check, ruff format --check, mypy --strict clean.
• codespell self-check passes (examples reuse the already-allowlisted wrod, so no new ignore-words entry).
• Full test_basic.py suite: 82 passed, 1 skipped.