Added merge-vs-per-file Word export on the scanned-PDF OCR tool — default stays one document; several per-file Word files download as a ZIP.

Date: 2026-09-08 22:55 Summary: Added merge-vs-per-file Word export on the scanned-PDF OCR tool — default stays one document; several per-file Word files download as a ZIP. Visibility: people

[question] 提供功能：分别按输入文件导出word还是合并在一个word导出

Tool links (English)

<batch-convert-scanned-pdf-to-word-with-ocr>

[try to solve]

Context

The shipped N4 page always packed every successful PDF into one DOCX. Users also need one Word file per input PDF without a second URL.

Process

1. Added a Word-export fieldset next to the language chips (not on the Convert/Download row): default “one Word for the batch”; optional “one Word per PDF”.
2. Refactored packing: buildDocxFromRows builds one DOCX from a list of files; packOutputs either merges or builds one DOCX per file. Two or more per-file successes load same-origin fflate and download a ZIP of .docx names derived from the PDF filenames. One success still uses Download Word.
3. Switching the radio or the page-image checkbox after OCR rebuilds the blob without re-running Tesseract.
4. How gained a fifth step; FAQ q10 explains the two modes. Ten-locale keys stayed in parity (105 keys).
5. verify:tool site/SEO/HTML smoke passed. Isolation needs CROSS_TOOL_UPDATE=1 only because unrelated neighbor files remain dirty in the working tree.

Root cause / analysis

Merge vs split is the same OCR job and the same queue. Splitting it into two URLs would be a doorway. The extra Information Gain is the export radio plus ZIP-when-several, with How labels matching Download Word / Download ZIP.

Solution

Keep the locked default (one Word). Per-file mode is a visible choice; the download button label follows the mode and the number of successful PDFs.

Notes / boundaries

• Caps unchanged: 5 PDFs, 20 pages, 12 MB.
• Per-file ZIP is Word files, not TXT.
• Live multi-file ZIP click was not run in a browser this session; prerendered HTML shows the radios and default Download Word.

[actions]

• Updated src/pages/batchConvertScannedPdfToWordWithOcrPage.ts and src/site/i18n/tools/batch-convert-scanned-pdf-to-word-with-ocr/*
• Updated work-tasks/batch-convert-scanned-pdf-to-word-with-ocr/02-tool-info.md
• npm run tool:touch -- --slug=batch-convert-scanned-pdf-to-word-with-ocr
• npm run verify:tool -- --slug=batch-convert-scanned-pdf-to-word-with-ocr (isolation re-run with CROSS_TOOL_UPDATE=1)