v0.9.7

Full Changelog: v0.9.6...v0.9.7

0.9.7 - 26-09-05

Added

• Added actions for unmatched references:
  • Open the DOI page when a DOI is available.
  • Search Google Scholar using the extracted title.
  • Copy the extracted title to the clipboard.
• Added support for headingless reference lists that use bare numeric markers without brackets or punctuation.

Improved

• Automatically rescan when newly rendered PDF pages are detected.
• Keep the manual scan button focused on scan state: it displays Scanning… while running and returns to Ref ↗ when complete.
• Display scan results only in the summary panel.
• Align the scan summary with the right edge of the reader.
• Better handle right-aligned one- and two-digit reference numbers in multi-column bibliographies.

Fixed

• Avoid treating middle initials in full author lists as title boundaries.
• Improved parsing of bare superscript-style AIP references.
• Prevent unmatched-reference action menus from triggering unnecessary rescans.

Verification

• Passed TypeScript type checking.
• Passed all 76 automated tests.
• Completed the production XPI build successfully.

v0.9.6

Full Changelog: v0.9.5...v0.9.6

0.9.6 - 2026-08-25

Added

• Highlight references that are not present in the Zotero library with a clearly visible gray background.
• Keep matched references yellow and clickable while leaving unmatched references non-interactive and without item badges.
• Support plain numbered reference markers such as 1. and 2) in addition to bracketed markers.

Fixed

• Parse MDPI-style numbered references even when PDF text extraction omits the space after the marker, such as 1.Mead.
• Avoid mistaking indented funding, data-availability, and acknowledgment blocks for a second text column.
• Continue numbered reference extraction onto later pages instead of stopping after the first 25 entries in the AppliedMath test paper.
• Preserve Roman-numbered subtitles such as : II. Multi-Coupled Systems when extracting citation titles.
• Resolve exact-title duplicates that share the same DOI or arXiv ID, preferring the Zotero item with a PDF attachment.
• Render links for plain numbered references after a citation has been matched.

v0.9.5

Full Changelog: v0.9.0...v0.9.5

0.9.5 - 2026-08-22

Fixed

• Reject partial-title matches, such as linking Principles of Vibration and Sound to Vibration and Sound.
• Distinguish book chapters from same-title books using Zotero item type and container metadata.
• Use author, year, item type, and container metadata to disambiguate items with identical titles.
• Leave ambiguous references unlinked instead of selecting an arbitrary library item.
• Prevent duplicate or overlapping PDF buttons on a single reference.
• Prefer PDF-bearing items when duplicate DOI or arXiv records exist.

Added

• Match title-less references using first author, year, journal, volume, and first page or article number.
• Support common astronomy, physics, mathematics, biomedical, and computer-science reference formats.
• Parse astronomy article numbers such as A70 and L13, organization authors, Vancouver-style names, comma-delimited mathematics titles, and repeated authors represented by an em dash.
• Normalize common journal names and abbreviations, including A&A, ApJ, ApJS, ARA&A, MNRAS, and Ann. Phys.

Improved

• Reconstruct two-column bibliographies using PDF text coordinates.
• Remove sequential margin line numbers from line-numbered preprints.
• Make fuzzy title matching more conservative by checking coverage, token order, and supporting metadata.
• Match and render references individually instead of searching the entire bibliography as one text block.

v0.9.0

Full Changelog: v0.8.2...v0.9.0

v0.8.2

Full Changelog:b464d54