Project Note JSON To Epub Agent Skill

專案筆記Json變成Epub

Use this skill when a project-note JSON or folder-derived JSON manuscript needs to become an EPUB, or an EPUB plus a PDF review/distribution edition, especially when the user wants chapter directories, index cards, cross-links, keyword numbering, versioned check editions, Chinese character counts, chapter-splitting plans, Kindle-clickable PDF links, or full EPUB/PDF link verification.

Chinese trigger phrases include:

• 專案筆記Json變成Epub
• 目錄索引卡交叉連結
• 把這章做成EPUB
• 章節拆書
• 關鍵字索引卡
• 雙向連結EPUB
• 同時生成PDF
• Kindle可以點擊連結

Core Workflow

• Inspect the JSON schema before generating anything.
• Identify title, sources, toc, chunks, index, and validation fields if present.
• Confirm each chunk has a stable code, title, body or source text, EPUB page/anchor, and Chinese character count.
• If the JSON is missing fields, derive them deterministically and write the enriched JSON first.

• Normalize the book identity.
• Use the user-provided book name exactly.
• For check editions, name versions as: {書名} 檢查版 YYYY-MM-DD NN.
• Use an ASCII-safe or underscore filename when needed: {書名}_檢查版_YYYY-MM-DD_NN.epub.
• Keep a simple download.epub copy for mobile sharing if the user needs an easy download link.

• Build the table of contents from project-note numbers.
• Preserve Luhmann/Zettelkasten-style codes such as 1.1, 3.3.c3b, and 3.3.d.1.
• Split into no more than five display levels: 部、章、節、項、目.
• Prefer original titles; generate compact titles only for unnamed chunks.
• Mark Chinese character counts at each level where possible.
• Treat the TOC as a directory card: every entry should preserve code, title, depth, page anchor, and future split-book key.

• Build the keyword index before EPUB rendering.
• Each keyword must have a stable id: K001, K002, ...
• Sort by descending weight, then keyword for deterministic order.
• Remove keywords below the requested threshold. If unspecified, ask or use the current project threshold.
• Each keyword may link to at most 5 target pages/chunks.
• Each index link needs display text, target, anchor, chunk code, title, and weight.
• Treat every keyword entry as an index card with id, keyword, weight, chapter attributes, selected page links, and cross-link metadata.

• Add bidirectional link data.
• Forward link: index keyword page number ->正文 chunk anchor.
• Backlink:正文 keyword marker -> index keyword anchor.
• Use the keyword id as backlink display text, such as K001.
• Store backlinks in JSON on the chunk, for example keyword_backlinks.
• Preserve cross-link attributes on each link: chapter key, section key, item key, selected page order, base weight, remote bonus, adjusted weight, and selection rank.

• Prepare chapter splitting data.
• Add a book_split_plan to JSON when the project is likely to become multiple chapter books.
• Each split entry should include split key, title, first/last code, first/last page, chunk count, CJK count, and output filename hint.
• Keep this data orthogonal to the EPUB render so chapters can later be exported independently.

• Render EPUB paths relative to the XHTML file that contains the link.
• From Text/index.xhtml to正文 parts, use part2.xhtml#p099, not Text/part2.xhtml#p099.
• From Text/part*.xhtml back to the index, use index.xhtml#idx-k001, not ../Text/index.xhtml#idx-k001.
• Keep canonical JSON targets if useful, but convert hrefs correctly during XHTML rendering.
• Put the keyword index card at the end of the book spine as book back matter: directory/intro first,正文 chunks next, index.xhtml last. This matches normal human book structure where an index appears after the main text.

• Generate a PDF companion when requested.
• Use the same enriched JSON, TOC, chunk anchors, keyword ids, and backlink data as the EPUB.
• Preserve the same human link model: directory/index entries jump to正文 anchors, and正文 keyword markers jump back to keyword index anchors.
• Use PDF internal destinations and link annotations, not external file URLs, so links remain clickable after sideloading to Kindle.
• Prefer visible blue linked text for PDF review copies because Kindle users need a clear tap target.
• Embed a CJK-capable font when generating Chinese PDFs. Avoid relying on a viewer's fallback fonts; black square glyphs are a build failure.
• Keep PDF anchors semantically aligned with EPUB anchors when possible: chunk p099 in EPUB should map to PDF destination p099, and keyword K001 should map to idx-k001.
• Treat PDF as a companion format; do not let PDF layout constraints mutate the canonical JSON or EPUB href rules.

• Validate before delivery.
• Open the EPUB zip and parse every XHTML file as XML.
• Collect all id anchors and all href links.
• Resolve every href relative to its source XHTML path.
• Fail the build if any target file or anchor is missing.
• Count expected forward links from index[].links.
• Count expected backlinks from chunks[].keyword_backlinks.
• Confirm actual EPUB forward/backlink counts match expected counts.
• Recalculate true_body_cjk_count from正文 body/chunk text only, excluding TOC, index cards, navigation files, metadata, CSS, generated link labels, and markup.
• Report the true body count separately from total EPUB or generated XHTML text counts.
• Confirm orthogonal_check.ok is true when the JSON includes directory/index/cross-link fields.
• For PDF output, inspect the PDF annotations with pypdf or equivalent and fail if internal GoTo/destination link counts do not match the expected forward/backlink model.
• Render the PDF to PNG and visually confirm Chinese text, blue link text, spacing, and page flow. If Chinese text renders as boxes, fix font embedding and rebuild.

EPUB Requirements

• EPUB mimetype must be the first zip entry and stored uncompressed.
• META-INF/container.xml, OEBPS/content.opf, OEBPS/nav.xhtml, and XHTML text files must exist.
• dc:title in content.opf must match the requested book/version title.
• The書末索引 must show keyword id, keyword, weight, and clickable page links.
• The OPF spine must place the keyword index page after all正文 pages unless the user explicitly requests a front-loaded index for review.

-正文 keyword markers should show the keyword plus a small clickable Kxxx backlink.

• Preserve readable styling; backlinks should not disrupt paragraph flow.

PDF Requirements

• Generate PDF only when requested, or when the user asks for Kindle-compatible clickable links.
• PDF links must be internal document links using named destinations or explicit GoTo actions.
• The書末索引 must include clickable links back to正文 chunk destinations.

-正文 keyword markers must include clickable Kxxx backlinks to the keyword index destination.

• PDF link counts should mirror EPUB expectations unless layout requires an explicitly documented exception.
• Use a CJK-capable embedded font for Chinese output. On macOS, a ReportLab build may use a TTC font such as /System/Library/Fonts/STHeiti Medium.ttc with a selected subfont index; in other environments choose an available Noto/Source Han CJK font.
• Render at least one representative page to PNG before delivery and inspect for missing glyphs, clipped text, overlapping links, and unreadable spacing.
• Kindle note: PDF internal links are the compatibility target. Do not depend on JavaScript, web URLs, or EPUB-only href behavior for the PDF edition.

Validation Checklist

Before telling the user it is finished, report:

errors: 0
title: ...
all_href_count: ...
index_forward_links_tested: ...
body_return_links_tested: ...
index_entries: ...
true_body_cjk_count: ...
file type: EPUB document
pdf_internal_links_tested: ...
pdf_render_checked: ...
pdf file type: PDF document

If any count fails, fix the generator and rebuild rather than editing EPUB output by hand.

Delivery

Return two links when possible:

• A simple mobile-safe link: download.epub
• The formal versioned EPUB filename
• The companion PDF filename when generated
• A validation report or short count summary when both EPUB and PDF are generated

Mention the version title and the link-test summary briefly.