Add Check for Reference-Only Changes #24

Implement an edit check that detects when a user has only added or modified references (content within <ref></ref> tags) without making any other changes to the article content.

Requirements

1. Reference-Only Detection

• Detect edits where the only changes are:
  • Adding new <ref></ref> tags with content
  • Modifying existing content within <ref></ref> tags
  • Adding or modifying self closed <ref /> tags
• No other text or content outside references should be changed

2. External Link Domain Verification

When a reference contains external links:

• Extract all domains from URLs in the added/modified references
• Check if each domain has been previously used in Wikipedia articles (namespace=0)
• Use Pywikibot's exturlusage() function for domain checking

Documentation:

• Pywikibot API: exturlusage
• MediaWiki API: Exturlusage

3. Auto-Approval Logic

Auto-approve when:

• ✅ Edit only adds/modifies references

Require manual review when:

• ✅ Edit only removes references (it is OK if it replaces one reference with another)

Implementation Notes

• Parse the diff to isolate changes within <ref></ref> tags
• Handle both self-closing <ref /> and standard <ref></ref> tags
• Refs can have name and group attributes: <ref name="..."> <ref group="...">
• There should be also SPAM link check for added URL:s but i will create separate issue ticket for this

Write unit tests covering:

• Adding a single reference
• Modifying an existing reference
• Adding multiple references
• Modifying multiple references
• Removing a reference (should not approve)
• Mixed edits with references and other content changes (should not approve)
• Adding self-closing reference tags
• References with name and group attributes

Acceptance Criteria

• Bot correctly identifies reference-only edits
• There is unit tests for the change