BookPrinting, #3

My basic method

Software:

• Acquire epub
• Use Pandoc to split epub to HTML
• delete unwanted chunks
• convert HTML to markdown
• clean file
• generate nice PDF using LaTeX header
• use bookbinder.js and print

Hardware:

• fold signatures, crease with bonefolder
• poke holes with awl
• sew
• cut out cardboard
• glue fabric
• complete case bound book

Answered questions:

• Why Epub?
  • Unless a PDF file is explicitly necessary, an Epub file is easier to reflow based on preference for page size, margins, text size, etc. Epub files are largely high quality and effectively make use of semantic markup.
• Why Pandoc?
  • Fast, free, easy.
• Why LaTeX?
  • LaTeX is used for high-quality typesetting. It is a free (as in free beer) software tool and can handle things like table of contents, footnotes, indices quite well.
• Why bookbinder.js ?
  • When a printed document is upwards of 5 pages in length, especially if it intended to be printed as a booklet, it is crucial to assemble a collection of "signatures" (smaller booklets) that are later joined in order to have a book that opens easily and turns pages well.

Unanswered questions

• How many pages and holes per signature?
  • Let's use 5 pages per signature and 6 holes.

commands being used

Epub -> HTML

pandoc input.epub -t chunkedhtml --chunk-template="%n_%h.html" -o chapters

for f in *.html; do pandoc "$f" -t markdown -o "${f%.html}.md"; done

Prune files... remove unnecessary html files

HTML -> markdown

for file in *.html; do
  pandoc "$file" -f html -t markdown --lua-filter=ignore-nav.lua -o "${file%.html}.md"
done

ignore-nav.lua

-- Helper function to check if a class list contains "sitenav"
local function has_sitenav_class(classes)
  for _, class in ipairs(classes) do
    if class == 'sitenav' then
      return true
    end
  end
  return false
end

-- Function to handle block containers (Div, Nav, Section, Header blocks)
function Div (el)
  if el.identifier == 'sitenav' or has_sitenav_class(el.classes) then
    return {} -- Strip the element completely
  end
end

-- Fallback for older Pandoc parsers handling generic block wrappers
function Block (el)
  if el.identifier == 'sitenav' or (el.classes and has_sitenav_class(el.classes)) then
    return {}
  end
end

Join all markdown files together as book.md . Minimal editing may be necessary after joining

markdown -> PDF (decide whether top-level-division is "part" or "chapter")

pandoc metadata.yaml book.md -H header.tex -o output.pdf --toc --pdf-engine=xelatex --top-level-division=part

metadata.yaml

---
title: "book's title"
author: "book's author"
documentclass: book
classoption:
  - twoside
  - openright
geometry:
  - a5paper
  - margin=20mm
mainfont: "Palatino Linotype"
---

header.tex

\usepackage{fancyhdr}
\usepackage{tocloft}
\usepackage{atbegshi}

% --- Blank page after title page ---
\AtBeginDocument{
  \maketitle
  \clearpage
  \thispagestyle{empty}
  \hbox{}
  \newpage
}

% --- Page Style Setup ---
\pagestyle{fancy}
\fancyhf{}
\fancyfoot[LE,RO]{\thepage} 
\renewcommand{\headrulewidth}{0pt}

\fancypagestyle{plain}{
  \fancyhf{}
  \fancyfoot[LE,RO]{\thepage}
  \renewcommand{\headrulewidth}{0pt}
}

\fancypagestyle{empty}{
  \fancyhf{}
}

% --- ToC Styling ---
\renewcommand{\cftchapleader}{\cftdotfill{\cftdotsep}}
\renewcommand{\cftpartleader}{\hfill}

% --- blank page after TOC --- 
\let\oldtableofcontents\tableofcontents
\renewcommand{\tableofcontents}{%
  \oldtableofcontents
  \clearpage
  \thispagestyle{empty}%
  \hbox{}%
  \clearpage
}

% --- no page numbers on blank pages ---
\makeatletter
\def\cleardoublepage{\clearpage\if@twoside \ifodd\c@page\else
  \hbox{}%
  \thispagestyle{empty}% <-- Forces the inkless style onto the skipped page
  \newpage
  \if@twocolumn\hbox{}\thispagestyle{empty}\newpage\fi
\fi\fi}
\makeatother

% --- Force a full blank spread before every Part ---
\let\oldpart\part
\renewcommand{\part}{
  \cleardoublepage    
  \thispagestyle{empty}
  \null              
  \newpage           
  \thispagestyle{empty}
  \newpage           
  \oldpart           
}