fix: match standard metadata names case-insensitively

[linjiaye929]

Summary

The public LXML scraper drops standard metadata such as <meta name="Description" content="Page summary">. HTML standard metadata names are ASCII case-insensitive, but the three current selectors require lowercase names.

Match description, keywords, and author names with XPath ASCII translation. Preserve content casing, first content-bearing tag order, missing/empty content behavior, and all Open Graph/Twitter handling. The legacy BeautifulSoup helper is outside this change.

Files and verification

• crawl4ai/utils.py: only the three standard-name predicates.
• tests/unit/test_metadata_name_case.py: public scraper regressions and compatibility controls.
• 118 focused/related tests pass, including 36 new cases (12 failed on unchanged develop).
• Local HTML fixtures; no browsers, live websites, or models. New tests pass Black and git diff --check; production formatting debt is pre-existing on both base and candidate. Full repository suite was not run. The exact-head Discord notification workflow completed with action_required; it requires maintainer action and does not establish hosted code-test success.

AI assistance

Developed and reviewed with AI assistance; no human full-diff review is claimed.

Checklist

• Based on and targets develop
• Self-reviewed the bounded change
• Added regressions and compatibility controls
• Existing metadata values and selection order preserved
• No new metadata fields or documentation contract introduced