PageScribe 051022bb4b OCR: fix paragraph duplication, caption and folio edge cases
- ParagraphReflow.toText re-appended the previous line on a plain line
  break (the line was already in the output), duplicating text across the
  spread. Only the join is now written, and a glued hyphen removes the
  hyphen that had already been emitted. Drops the now-unused join().
- CaptionFilter: the gapBelow branches were Int/Double and inferred
  Comparable<*>; widen to Double.
- PageNumberDetector: tighten folio geometry and margin isolation.

Add the 83-test JVM/Robolectric suite that locks these down.
2026-10-15 10:20:29 -07:00
Description
No description provided
Apache-2.0 1.5 MiB
Languages
Kotlin 92%
Python 8%