fix: update CompanyScraper selectors for new LinkedIn DOM

[choru-k]

Summary

LinkedIn changed their company page DOM structure, causing CompanyScraper to return null for all fields except linkedin_url and name. This PR updates the selectors to match the new structure.

Changes:

• Navigate to /about page for more detailed company info
• Update CSS selectors for about section (section.org-about-module__margin-bottom p.break-words)
• Refactor overview extraction into helper methods:
  • _parse_dl_definition_list() - Parses dl.overflow-hidden dt/dd structure
  • _parse_top_card_info() - Parses .org-top-card-summary-info-list__info-item
  • _find_website_link() - Extracts company website URL
• Add multi-language support for field detection (English, Korean, Japanese)
• Infer field types from value patterns rather than relying solely on localized labels

Test Results

Tested with multiple companies:

# ✅ get_company_profile('anthropicresearch')
{
  'name': 'Anthropic',
  'about_us': 'Anthropic is an AI safety company...',
  'website': 'https://www.anthropic.com',
  'industry': 'Research Services',
  'company_size': '501-1,000 employees',
  ...
}

# ✅ get_company_profile('microsoft')
{
  'name': 'Microsoft',
  'about_us': 'Every company has a mission...',
  'website': 'https://news.microsoft.com',
  'industry': 'Software Development',
  'company_size': '10,001+ employees',
  'headquarters': 'Redmond, Washington',
  ...
}

Backwards Compatibility

The changes maintain backwards compatibility:

• Still tries the old dt/dd selectors as fallback
• Still parses .org-top-card-summary-info-list__info-item if the new structure isn't found
• Returns the same Company object structure

Fixes #238