1.3.0

[1.3.0] - 2025-08-05

Added

• PDF Viewer Integration:

  • Added a new PDF viewer route (/viewer) that allows users to open a PDF document at a specific page using ?path=...&page=....
  • Integrated PDF.js from Mozilla to render PDF pages directly in the browser using a <canvas> and a dynamic text layer.
  • Implemented search term highlighting for a given query using ?q=..., applied on the specified page.
  • Highlighting is case-insensitive and styled using <mark> elements injected into the text layer.
  • The highlight feature now retrieves terms directly from the Elasticsearch results (search highlights), enabling a more seamless experience when navigating between results.
  • Added parsing and injection of the highlighted terms into the PDF viewer dynamically, improving the user experience.
  • Limitations: Currently highlights only the first occurrence of the search term per span (this limitation will be improved in future versions).
  • Improved the highlight feature to correctly show all matches including those with accented characters, fixing issues where accented terms were partially or incorrectly highlighted.

• Project Management:

  • Added TODO.md document to track pending features, improvements, and technical debt.
  • Serves as a lightweight roadmap for contributors and team members.

• Accent and Special Character Normalization:

  • The indexer now replaces accented characters and special variations of vowels (e.g., á, é, í, ó, ú, ü) with their plain equivalents (a, e, i, o, u) during indexing.
  • This improves the consistency of search queries and results when users omit accents.

Changed

• Indexer and Search Refactoring:

  • Major refactor of the indexer and search logic to improve maintainability and search consistency.
  • Normalized input during both indexing and querying phases to better handle special characters and improve match accuracy.

• Search Results Handling:

  • The highlight terms fetched from Elasticsearch are now processed and passed to the PDF viewer for more accurate highlighting.
  • The search terms in Elasticsearch are parsed to ensure they are appropriately reflected in the PDF viewer.
  • Enhanced handling of search results to integrate smoothly with the PDF viewer.

Known Issues

• Character Encoding:
  • There may be issues with certain characters (e.g., accented characters) not being properly highlighted in the PDF viewer. This issue will be addressed in future versions.
  • The rendering of accented characters such as José in the highlights might not be perfect due to encoding differences between the PDF content and the search terms.

What's Changed

• Feat/pdfjs highlight by @josego85 in #1

New Contributors

• @josego85 made their first contribution in #1

Full Changelog: v1.2.2...v1.3.0