Ground truth for report-text.pdf (YYSI synthetic demo, 2026-11-11). All amounts are invented. Logical reading order below; note the two-column sections are column-wise here, while row-wise extraction interleaves them. === Page 1 === Quarterly Field Report — Q3 2026 Demo document generated by YYSI for the PDF extraction guide. All data is synthetic. This page has a real text layer. Open it in any PDF reader and you can select these words with the mouse and copy them. Extraction tools read the same text without OCR — that is why the extracted result keeps every word. Left column: maintenance A. Filter replaced on 12 August. B. Coolant topped up, no leaks. C. Tyre pressure set to 2.4 bar. Right column: inspections D. Brake pads measured at 6 mm. E. Lights and horn tested, pass. F. Next visit booked 30 October. Item Qty Unit price Amount Air filter 2 14.50 29.00 Wiper blades 4 9.80 39.20 Coolant (1 L) 3 12.35 37.05 Total 105.25 === Page 2 === 库存盘点示例(合成数据) 本页带有真实文字层:在阅读器里可以直接选中并复制这些汉字, 提取工具不需要 OCR 也能读到同样的内容。 但换行位置和先后顺序取决于 PDF 内部书写顺序,双栏内容尤其要核对。 左栏:保养记录 甲.已更换机油滤芯。 乙.已补充玻璃水。 右栏:安全检查 丙.灯光测试通过。 丁.胎压恢复正常。 品名 数量 单价(元) 金额(元) 空气滤芯 2 14.50 29.00 雨刮片 4 9.80 39.20 冷却液(1升) 3 12.35 37.05 合计 105.25