From web page to printable PDF for reading later

Rather than using a normal-person's "Read Later" service, I print long-form web articles for reading later. I print them, pile them up, and read them all when I have some time away from the computer. It's the only way I can truly pay attention to them.

The process took some work to dial in, but I've gotten it close to how I like it. It goes like this:

1. Save the page as Markdown using the Obsidian Web Clipper
2. Convert the Markdown to PDF using Pandoc
3. Print!

First, the Obsidian Web Clipper. I have a custom "template" configured for this. The template doesn't use Properties. I put the entire block of front matter into the Content field, using variables to insert the specifics, so no need to worry about the way the clipper renders properties...

---
title: "{{title}}"
author: "{{author|safe_name}}"
source: {{url}}
created: {{date}}
published: {{published|date:"YYYY-MM-DD"}}
---

> ## Excerpt
> {{description}}

{{content}}

A copy of the template for importing directly into the web clipper is here

The magic comes from Pandoc. I convert the Markdown to PDF via LaTeX. The default.latex template that ships with Pandoc works well. It uses all sorts of variables that can be passed in via YAML front matter or via a defaults file. I only just learned about the default file option, so that cleaned up the front matter dramatically. Here's my ~/.pandoc/defaults/article.yaml defaults file:

pdf-engine: xelatex
template: ~/.pandoc/templates/default.latex

variables:
  documentclass: scrartcl
  mainfont: XCharter
  sansfont: Playfair Display
  linestretch: 1.15
  header-includes:
    - \setkomafont{title}{\sffamily\bfseries}
    - \setkomafont{section}{\sffamily\bfseries}
    - \setkomafont{subsection}{\sffamily\bfseries}
  classoption:
    - DIV=14
    - twocolumn 

This sets fonts, sizes, document class, etc. The DIV=14 is new to me. With KOMA classes, it sets the text width, taking font size into account. This is easier than including specific geometry. Smaller numbers = narrower text. Drop the twocolumn line to print full-width.

I could create separate default files for different layouts and styles. Use the -d option, e.g. pandoc -d article. I might make one for use on the Remarkable tablet, which likes a larger font and wider body.

A shell script, md2pdf.sh does the conversion for me. The script used to be about 10 lines of sloppy bash with a bunch of stuff hard-coded. I had Claude help me make it more robust, and here's the latest version:

#!/bin/sh
set -eu

DEFAULT_DEFAULTS_FILE="$HOME/.pandoc/defaults/article.yaml"

usage() {
    cat >&2 <<EOF
Usage: $(basename "$0") [OPTIONS] FILE

Convert a Markdown file to PDF with pandoc and open it.

Options:
  -o DIR    Output directory (default: same directory as FILE)
  -d FILE   Pandoc defaults file (default: $DEFAULT_DEFAULTS_FILE)
  -n        Don't open the PDF after creating it
  -h        Show this help
EOF
    exit 1
}

output_dir=""
defaults_file="$DEFAULT_DEFAULTS_FILE"
open_after=1

while getopts "o:d:nh" opt; do
    case "$opt" in
        o) output_dir="$OPTARG" ;;
        d) defaults_file="$OPTARG" ;;
        n) open_after=0 ;;
        h) usage ;;
        *) usage ;;
    esac
done
shift $((OPTIND - 1))

[ $# -ge 1 ] || usage

input="$1"

if [ ! -f "$input" ]; then
    echo "Error: input file '$input' not found" >&2
    exit 1
fi

if [ ! -f "$defaults_file" ]; then
    echo "Error: defaults file '$defaults_file' not found" >&2
    exit 1
fi

# Default output dir to source dir
[ -n "$output_dir" ] || output_dir=$(dirname "$input")
mkdir -p "$output_dir"

stem=$(basename "$input")
stem="${stem%.*}"
output="$output_dir/$stem.pdf"

# Capture pandoc stderr so we can report it clearly on failure
log=$(mktemp)
trap 'rm -f "$log"' EXIT

if ! pandoc --defaults "$defaults_file" "$input" -o "$output" 2>"$log"; then
    echo "pandoc failed converting '$input'" >&2
    echo "----- pandoc output -----" >&2
    cat "$log" >&2
    echo "-------------------------" >&2
    echo "Command: pandoc --defaults '$defaults_file' '$input' -o '$output'" >&2
    exit 1
fi

# Surface warnings even on a successful run
[ -s "$log" ] && cat "$log" >&2

echo "Created: $output"

if [ "$open_after" -eq 1 ]; then
    xdg-open "$output" >/dev/null 2>&1 &
fi

So the simple version (assuming md2pdf.sh is in your path) is:

md2pdf.sh my-article.md

It's not quite cross-platform (e.g xdg-open on Linux vs open on macOS) but that shouldn't be to difficult to do.

This seems like a lot, but once it's in place it's a couple of clicks and a quick shell command.

Let me know if there's anything unclear or incorrect here. Or if you have suggestions for improvements.