Changelog

[1.1.0]

🌐 New Feature

• Context-Aware Linking

  • Introduced a sophisticated, context-aware phrase linking engine that intelligently detects and links entire multi-word phrases (e.g., “French Revolution”, “Public Administration Reforms”) within context.
  • Queries are evaluated based on sequential token alignment, ensuring phrase integrity and semantic clarity.
  • A tiered scoring heuristic (rule-based evaluation) ranks link suggestions by contextual relevance:

    • Exact match (1.0) – phrase and target are identical.
    • Starts-with match (0.8) – target begins with the phrase.
    • Contains match (0.5) – phrase appears anywhere in the target.

🚀 Enhancements

• Refined trigger detection
  • Entire phrase beneath the cursor—including interspersed whitespace—is now captured and processed holistically.
  • Employs sentence-boundary parsing to improve resilience (ability to handle edge cases) and consistency across writing styles.

🐛 Bug Fixes

• Fixed multi-word linking regression
  • Previously, only the final token in a phrase was linked (e.g., [[French Revolution]] linked only “Revolution”).
  • Now, entire phrases are correctly linked, maintaining semantic cohesion (meaning-preserving structure).

📈 Retained Capabilities

• Intelligent single-word linking
• Auto-linking for dates and timestamps
• Support for heading, block, and tag insertion