ExtractPDF4J/ExtractPDF4J

Header-anchored column alignment across pages

开放

#7 创建于 2025年8月25日

 (2 条评论) (0 个反应) (2 位负责人)Java (44 个派生)auto 404
enhancementgood first issue

仓库指标

星标
 (545 个星标)
PR 合并指标
 (30 天内没有已合并 PR)

描述

Why

Columns drift when layouts change per page. We need to anchor to headers and realign.

Tasks

  • Detect header row per page (fuzzy match against canonical header set)
  • Map per-page headers → canonical order
  • Snap column bounds to matched headers (with tolerance)
  • Add unit tests + 2 real PDFs as fixtures

Acceptance Criteria

  • Multi-page PDFs with shifting headers produce aligned CSVs (no column swaps).

贡献者指南