[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-78065":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":14,"forks30d":14,"starsTrendScore":18,"compositeScore":19,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":14,"starSnapshotCount":14,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},78065,"public-domain-books-translation","SaberOnGo\u002Fpublic-domain-books-translation","SaberOnGo","Public domain books translation project",null,"Python",205,20,1,0,7,18,179,21,3.97,false,"main",true,[],"2026-06-12 02:03:46","# LifeBook Shufang: Global Public-Domain Book Translation and EPUB Collaboration\n\n\u003Ctable align=\"center\">\n  \u003Ctr>\n    \u003Ctd align=\"center\">\u003Ch3>\u003Ca href=\".\u002FREADME.zh-CN.md\">简体中文\u003C\u002Fa>\u003C\u002Fh3>\u003C\u002Ftd>\n    \u003Ctd align=\"center\">\u003Ch3>\u003Ca href=\".\u002Freadme\u002FREADME.zh-TW.md\">繁體中文\u003C\u002Fa>\u003C\u002Fh3>\u003C\u002Ftd>\n    \u003Ctd align=\"center\">\u003Ch3>\u003Ca href=\".\u002FREADME.md\">English\u003C\u002Fa>\u003C\u002Fh3>\u003C\u002Ftd>\n    \u003Ctd align=\"center\">\u003Ch3>\u003Ca href=\".\u002Freadme\u002FREADME.ja.md\">日本語\u003C\u002Fa>\u003C\u002Fh3>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\nLifeBook Shufang is a global, multilingual collaboration project for translating public-domain books and producing readable EPUB editions.\n\nThe goal is not to publish raw AI output. AI can help with repetitive work: finding and cleaning source text, splitting chapters, drafting translations, creating glossaries, checking omissions, and building EPUB files. But good books still need human judgment. People are needed to read, question, correct, polish, verify names and terms, and test the final book in real reading apps.\n\nYou do not need to be a programmer to help. Reading a chapter and saying \"this part is confusing\" is already useful.\n\n## What We Are Trying To Do\n\n- Choose public-domain books from reliable sources such as Project Gutenberg, Wikisource, and Standard Ebooks.\n- Preserve source and rights evidence before translation.\n- Avoid modern copyrighted editions, modern annotations, and unclear download sources.\n- Use AI for traceable drafting and checking, not for one-shot unreviewed publishing.\n- Keep source chapters, draft translations, final chapters, review files, glossary files, and EPUB output in a structure that other people can inspect.\n- Let contributors help with small tasks: reading, proofreading, terminology, source research, layout checks, and EPUB testing.\n\nThis is intentionally modest. Think of it as a small book workshop: someone finds a book, someone checks the source, someone reviews a chapter, someone tests the EPUB. Many small contributions can make one book better.\n\n## Easiest Way To Start\n\nMost contributors do not need to manually copy folders or fill in metadata files. Give an AI assistant the template folder, a public-domain source URL if you have one, and the translation direction. The AI should create the book project, fill in metadata, document source evidence, run the translation pipeline, and build the EPUB.\n\nUsually you only need to provide:\n\n- The book you want to work on.\n- A public-domain source URL, if known.\n- The translation direction, such as English to Spanish, French to Japanese, Chinese to English, or Japanese to German.\n- The existing language-pair template to use. If that template does not exist yet, add the matching template directory first.\n\nExample prompt. This uses French to English as a concrete direction. If `fr-en` does not exist yet, add that language-pair template first:\n\n```text\n\u002Fgoal Create an English EPUB from a French public-domain book.\nFetch the source text from {reliable public-domain source URL}.\nUse template\u002Fepub_pipeline\u002Fcommon plus template\u002Fepub_pipeline\u002Ffr-en, and let the orchestrator prompt run the full workflow:\nsource and rights review, book research, pre-translation trials, chapter translation, chapter review,\nchapter gates, preproduction stage 1, sample EPUB review, full EPUB production, post-EPUB stratified random spot-check, independent review,\nrevision routing, versioned release, final output, and retrospective.\nGenerate output\u002Fbook.epub, pass epubcheck, and create output\u002Frelease\u002Fbook_v0.0.1.epub with a bilingual release note.\n```\n\nIf you only know a title, ask AI to find a reliable public-domain source first:\n\n```text\nPlease find a reliable public-domain source for {book title}.\nPrioritize Project Gutenberg, Wikisource, and Standard Ebooks.\nAfter source and rights risks are checked, choose the matching language-pair template. For French to English, use or add template\u002Fepub_pipeline\u002Ffr-en; for Japanese to Spanish, use or add template\u002Fepub_pipeline\u002Fja-es; for English to Simplified Chinese, use template\u002Fepub_pipeline\u002Fen-zh-Hans. Then create the new book project with books\u002Fscripts\u002Fcreate_book_project.py; it copies template\u002Fepub_pipeline\u002Fcommon first, overlays the language-pair template, and places the project under books\u002F{target}\u002F{number}_{book_slug}\u002F.\n```\n\n## Repository Structure\n\n### `template\u002Fepub_pipeline\u002F`\n\nThis is the reusable book-production template area. `common\u002F` contains shared EPUB workflow contracts, rights checks, state files, scripts, and production rules. Each language-pair directory contains prompts, glossary\u002Fstyle guidance, typography expectations, and review rules for that direction. For a new book, ask AI to use books\u002Fscripts\u002Fcreate_book_project.py so the project is created under books\u002F{target}\u002F{number}_{book_slug}\u002F with the next numeric prefix; do not put real book data into the template folder itself.\n\nImportant files and folders:\n\n- `template\u002Fepub_pipeline\u002FREADME.md`: template layout guide.\n- `template\u002Fepub_pipeline\u002Fcommon\u002FPIPELINE_SPEC.md`: pipeline contract and directory rules.\n- `template\u002Fepub_pipeline\u002Ftargets\u002Fzh-Hans\u002Fquality_framework\u002FREADME.md`: Simplified Chinese target-language quality framework.\n- `template\u002Fepub_pipeline\u002Fen-zh-Hans\u002FREADME.md`: one currently available language-pair template guide.\n- `template\u002Fepub_pipeline\u002Fen-zh-Hans\u002FMASTER_PROMPT.md`: master prompt for starting a new book project.\n- `prompts\u002F`: step-by-step prompts from ingestion to review, EPUB production, and retrospective.\n- `metadata\u002F`: book metadata, rights checklist, source evidence, style profile.\n- `chapters\u002F`: source chapters, translated drafts, and final chapters.\n- `qa\u002F`: fidelity, readability, terminology, imagery, and gate reviews.\n- `preproduction\u002F`: cover, metadata, layout, and sample EPUB checks.\n- `reviews\u002F`: independent review and scorecards.\n- `output\u002F`: current build EPUB, validation results, and versioned release artifacts under `output\u002Frelease\u002F`.\n\n### `books\u002F`\n\nThis folder contains actual book projects grouped by target language. New projects use this layout:\n\n```text\nbooks\u002F{target}\u002F{number}_{book_slug}\u002F\n```\n\nFor example, a Simplified Chinese edition can live at `books\u002Fzh-Hans\u002F1_pg20923_a_negro_explorer_at_the_north_pole\u002F`. That sample is based on Project Gutenberg #20923, Matthew A. Henson's *A Negro Explorer at the North Pole*. It includes source evidence, rights notes, 26 source chapters, translated chapters, final chapters, review files, generated EPUB output, and an EPUBCheck report with zero fatal errors, zero errors, and zero warnings.\n\nThe example proves that the workflow can run end to end. It does not mean every chapter has already received final human editorial approval.\n\nNode.js tooling is shared at `books\u002F`: run `npm install` once from `books\u002F`, then run `npm run ...` inside each book project. Do not create a duplicate `node_modules\u002F` inside every book directory unless a specific book records a justified exception. Book-local scripts must walk up to `books\u002Fnode_modules\u002F` because book projects may be nested under target-language directories.\n\n### `template\u002Fepub_pipeline\u002Ftargets\u002F`\n\nThis folder stores target-language quality frameworks. For example, `template\u002Fepub_pipeline\u002Ftargets\u002Fzh-Hans\u002Fquality_framework\u002F` contains Simplified Chinese translation quality standards, trial-translation rules, benchmark policy, and chapter gates.\n\n### `doc\u002Fpublic\u002F`\n\nThis folder stores public notes for candidate books, copyright screening, and source research.\n\n### `research\u002F`\n\nThis folder stores language-pair-specific research artifacts and historical exploration files. For example, `research\u002Fen-zh-Hans\u002Fbook-discovery\u002F` contains English-to-Simplified-Chinese candidate-book audits and scripts. Files in this area are not global defaults; other language pairs should create their own research folders.\n\n## How You Can Help\n\nUseful contributions include:\n\n- Read a few pages and report awkward passages.\n- Compare one source paragraph with its translation.\n- Check names, places, and recurring terms.\n- Find typos or punctuation issues.\n- Test an EPUB on a phone, tablet, or e-reader.\n- Research whether a candidate book is truly public domain.\n- Improve cover, metadata, layout, or EPUB compatibility.\n- Review whether a chapter sounds like natural Chinese or natural English.\n\nSimple feedback format:\n\n```text\nBook:\nChapter:\nLocation or sentence:\nProblem:\nSuggestion, if any:\n```\n\n## Human Checkpoints\n\nAI can run most of the workflow automatically, but people are especially useful at these points:\n\n- Book research: `metadata\u002Fbook_specific_translation_research.md`, `metadata\u002Fstyle_profile.md`, `glossary\u002Fterms.csv`.\n- Pre-translation trial: `qa\u002Fpretranslation\u002Fpretranslation_report.md`.\n- Chapter control and review: `chapters\u002Fsrc\u002F`, `chapters\u002Ftranslated\u002F`, `chapters\u002Ffinal\u002F`, `qa\u002Fchapter_controls\u002F`, `qa\u002Ffidelity\u002F`, `qa\u002Freadability\u002F`, `qa\u002Fterminology\u002F`, `qa\u002Fimagery\u002F`, `qa\u002Fgates\u002F`.\n- Preproduction stage 1: `preproduction\u002Fstage1\u002Fproduction_spec.md`.\n- Preproduction stage 2: `preproduction\u002Fstage2_sample\u002Fsample_book.epub` and `sample_review.md`.\n- Final output: `output\u002Fbook.epub`, `output\u002Fepubcheck.json`, `reviews\u002F`, `reviews\u002Frandom_spotcheck\u002Fround_XXX\u002F` for human-checkable stratified samples, evidence, fixes, and closure records, plus `output\u002Frelease\u002Fbook_vX.X.X.epub` and `release_note_vX.X.X.md`.\n\n## Translation Directions\n\nTemplates are organized by concrete translation direction. `common` is shared by all language pairs. Directories such as `en-zh-Hans`, `fr-en`, `ja-es`, and `zh-Hant-de` identify the source and target languages.\n\nThe project can support concrete directions such as:\n\n- English public-domain novels into Spanish.\n- French essays into Japanese.\n- Chinese public-domain works into English.\n- Japanese travel writing into German.\n- German philosophy into Traditional Chinese.\n- Arabic historical texts into Indonesian.\n\nFor every direction, the same principles apply: verify rights, document sources, research before translating, run sample trials first, review chapters carefully, and validate the final EPUB.\n\n## Rules We Care About\n\n- Do not use modern copyrighted translations as source material.\n- Do not use pirate websites or unclear EPUB downloads.\n- Do not treat raw AI output as publishable text.\n- Do not write real book data into the template folder.\n- Keep review records and failure records.\n- Prefer small, reviewable improvements over untraceable whole-book rewrites.\n\n## License And Rights Notes\n\nEach source book must be checked separately. A text may be public domain in one country but not automatically public domain everywhere.\n\nProject Gutenberg texts are often public domain in the United States. Before wider distribution, contributors should review the copyright status for the target region.\n\nTranslations, notes, covers, formatting, EPUB packaging, and other non-code content produced by this project are released to the public under `CC BY-NC-SA 4.0` by default. Third-party commercial use requires separate permission from LifeBook Shufang and the relevant rights holders.\n\nBy contributing, contributors agree that their contributions may be included in this project for public release and may be used by LifeBook Shufang in LifeBook products and services. LifeBook Shufang is responsible for project organization, publication, quality control, license management, and contributor return arrangements. Specific attribution and return methods are determined according to contribution status and project rules.\n\nSee:\n\n- [LICENSE.en.md](.\u002Flicense\u002FLICENSE.en.md): public content license and code license.\n- [CONTRIBUTING.en.md](.\u002Flicense\u002FCONTRIBUTING.en.md): contributor authorization and participation rules.\n- [COMMERCIAL_LICENSE.en.md](.\u002Flicense\u002FCOMMERCIAL_LICENSE.en.md): third-party commercial use notes.\n\n## A Practical Invitation\n\nIf you like books and want to help bring overlooked public-domain works to more readers, you can start small. Read a chapter. Mark one awkward sentence. Check one name. Test one EPUB. A book improves through many small acts of attention.\n","LifeBook Shufang 是一个全球多语言协作项目，旨在翻译公有领域的书籍并生成可读的EPUB版本。该项目利用AI技术处理重复性工作，如查找和清理源文本、拆分章节、草拟翻译、创建词汇表、检查遗漏以及构建EPUB文件，但最终仍需人工审核与校对以确保质量。适合希望参与文学作品翻译、审校或测试EPUB文件格式的志愿者加入，无需编程背景即可贡献自己的力量。",2,"2026-06-11 03:56:25","CREATED_QUERY"]