Input (PDF/DOCX/EPUB)
  │
  ▼
Calibre ebook-convert → HTMLZ → HTML → Markdown
  │
  ▼
Split into chunks (~6000 chars each)
  │  manifest.json tracks SHA-256 hashes
  ▼
Parallel subagents (8 concurrent by default)
  │  each: read chunk → translate → write output_chunk*.md
  ▼
Validate (manifest hash check, 1:1 source↔output match)
  │
  ▼
Merge → Pandoc → HTML (with TOC) → Calibre → DOCX / EPUB / PDF

# 1. Calibre (provides ebook-convert)
# macOS
brew install --cask calibre
# Linux
sudo apt-get install calibre
# Or download from https://calibre-ebook.com/

# 2. Pandoc
brew install pandoc        # macOS
sudo apt-get install pandoc # Linux

# 3. Python dependencies
pip install pypandoc beautifulsoup4

ebook-convert --version
pandoc --version
python3 -c "import pypandoc; print('pypandoc ok')"

npx skills add deusyu/translate-book -a claude-code -g

clawhub install translate-book

git clone https://github.com/deusyu/translate-book.git ~/.claude/skills/translate-book

translate /path/to/book.pdf to Chinese

translate ~/Downloads/mybook.epub to Japanese

/translate-book translate /path/to/book.docx to French

python3 scripts/convert.py /path/to/book.pdf --olang zh

# For EPUB input
python3 scripts/convert.py /path/to/book.epub --olang ja

# For DOCX input
python3 scripts/convert.py /path/to/book.docx --olang fr

# Each subagent receives exactly this task:
Read chunk0042.md → translate to target language → write output_chunk0042.md

python3 scripts/merge_and_build.py \
  --temp-dir book_name_temp \
  --title "《Book Title in Target Language》"

translate-book/
├── SKILL.md                    # Claude Code skill definition (orchestrator)
├── scripts/
│   ├── convert.py              # PDF/DOCX/EPUB → Markdown chunks via Calibre HTMLZ
│   ├── manifest.py             # SHA-256 chunk tracking and merge validation
│   ├── merge_and_build.py      # Merge chunks → HTML → DOCX/EPUB/PDF
│   ├── calibre_html_publish.py # Calibre wrapper for format conversion
│   ├── template.html           # Web HTML template with floating TOC
│   └── template_ebook.html     # Ebook HTML template
└── README.md

# scripts/manifest.py (conceptual usage)

# During convert.py — records source hashes
manifest = {
    "chunk0001.md": "sha256:abc123...",
    "chunk0002.md": "sha256:def456...",
    # ...
}

# During merge_and_build.py — validates before merging
# 1. Check every chunk has a corresponding output_chunk
# 2. Re-hash source chunks and compare against manifest
# 3. Reject if any hash mismatches (stale/corrupt output)
# 4. Reject if any output file is empty

# 1. Install the skill
npx skills add deusyu/translate-book -a claude-code -g

# 2. Open Claude Code in your working directory
cd ~/books

# 3. Say in Claude Code:
# "translate clean-code.pdf to Chinese"

# Claude Code will:
# - Run convert.py to split into chunks
# - Launch 8 parallel subagents per batch
# - Each subagent translates one chunk
# - Validate all outputs via manifest
# - Merge and build all formats

# 4. Outputs appear in:
ls clean-code_temp/
# chunk0001.md  chunk0002.md  ...  (source)
# output_chunk0001.md  ...         (translated)
# manifest.json
# output.md
# book.html
# book.docx
# book.epub
# book.pdf

# If translation is interrupted, just re-run the same command:
# "translate clean-code.pdf to Chinese"

# The skill detects existing output_chunk*.md files
# and skips already-translated chunks automatically.
# Only missing or failed chunks are retried.

# Delete only the final artifacts (keeps translated chunks)
cd book_name_temp/
rm -f output.md book*.html book.docx book.epub book.pdf

# Re-run merge step
python3 ../scripts/merge_and_build.py \
  --temp-dir . \
  --title "《New Title》"

# Check which chunks are missing translation
ls book_temp/chunk*.md | wc -l          # total source chunks
ls book_temp/output_chunk*.md | wc -l   # translated chunks so far

# Find missing output chunks
for f in book_temp/chunk*.md; do
  base=$(basename "$f" .md)
  out="book_temp/output_${base}.md"
  if [ ! -f "$out" ] || [ ! -s "$out" ]; then
    echo "Missing: $out"
  fi
done

# Check manifest
cat book_temp/manifest.json | python3 -m json.tool | head -30