extract_mermaid.sh

#!/usr/bin/env bash
# Extracts each ```mermaid fence from a Markdown file and renders it as a
# standalone image file. Companion to md_to_pdf.sh (same toolchain, same
# install steps) for when you want just the figure, not a full-document PDF
# that can split a tall diagram across pages.
#
# Usage: ./extract_mermaid.sh <input.md> [svg|png|pdf|both] [outdir]
#   both -> svg AND png (default): svg is the vector render kept beside the
#           .md (PDF pipeline, GitHub preview); png (3x) is what a pandoc
#           docx build must embed. A pandoc-built .docx with a bare svg and
#           no rsvg-convert fallback shows EMPTY FRAMES in Word, so any
#           document destined for .docx needs the png twin.
#   svg  -> vector only; fine for manual Insert > Pictures in Word 365,
#           NOT sufficient for pandoc docx builds (see above)
#   png  -> raster only at 3x scale; for slides or chat
#   pdf  -> one diagram per file, scaled to fit a single page (--pdfFit)
#
# Output naming: <input-basename>-1.<fmt>, -2, ... (one per mermaid fence,
# in document order). Spaces in the input name become underscores.
#
# v3 changes (August 2026):
#   - `both` mode added and made the default. Discovered while building the
#     Authorization-Derived Topology docx: pandoc embeds a bare svg with no
#     raster fallback when rsvg-convert is absent, and Word renders those as
#     blank frames. Rendering svg+png together makes the docx-safe asset
#     exist by default instead of being a rescue step.
#
# v2: ${EXTRA[@]+...} guard — macOS ships bash 3.2, where expanding an empty
# array under `set -u` is an "unbound variable" error (fixed upstream in 4.4).

set -euo pipefail

[[ $# -lt 1 ]] && { echo "Usage: $0 <input.md> [svg|png|pdf|both] [outdir]"; exit 1; }
[[ -f "$1" ]] || { echo "Error: $1 not found"; exit 1; }

INPUT="$1"
FMT="${2:-both}"
case "$FMT" in svg|png|pdf) FORMATS="$FMT";; both) FORMATS="svg png";; *) echo "Error: format must be svg, png, pdf, or both"; exit 1;; esac
OUTDIR_RAW="${3:-.}"
mkdir -p "$OUTDIR_RAW"
OUTDIR="$(cd "$OUTDIR_RAW" && pwd)"

CHROME="$(ls -d "$HOME/.cache/puppeteer/chrome-headless-shell"/*/chrome-headless-shell-mac-arm64/chrome-headless-shell 2>/dev/null | sort -V | tail -1)"

command -v mmdc &>/dev/null || { echo "Error: mmdc not found. Run: npm install -g @mermaid-js/mermaid-cli"; exit 1; }
[[ -x "${CHROME:-}" ]] || { echo "Error: chrome-headless-shell not found. Run: npx puppeteer browsers install chrome-headless-shell"; exit 1; }

WORK_DIR="$(mktemp -d)"
trap 'rm -rf "$WORK_DIR"' EXIT

# spaces -> '_' so mmdc's generated filenames stay predictable
BASENAME="$(basename "${INPUT%.md}" | tr ' ' '_')"
PUPPETEER_CFG="$WORK_DIR/puppeteer_config.json"
printf '{"executablePath": "%s", "args": ["--no-sandbox"]}\n' "$CHROME" > "$PUPPETEER_CFG"

shopt -s nullglob
TOTAL=0

for F in $FORMATS; do
  EXTRA=()
  [[ "$F" == "png" ]] && EXTRA+=(--scale 3)
  [[ "$F" == "pdf" ]] && EXTRA+=(--pdfFit)

  echo "-> Rendering Mermaid fences from $(basename "$INPUT") as $F..."
  mmdc \
    -i "$INPUT" \
    -o "$WORK_DIR/${BASENAME}.md" \
    -e "$F" \
    -b white \
    -p "$PUPPETEER_CFG" \
    ${EXTRA[@]+"${EXTRA[@]}"} \
    --quiet

  FILES=( "$WORK_DIR/${BASENAME}"-*."$F" )
  [[ ${#FILES[@]} -eq 0 ]] && { echo "Error: no \`\`\`mermaid fences found in $INPUT"; exit 1; }

  for f in "${FILES[@]}"; do
    cp "$f" "$OUTDIR/"
    echo "OK Saved: $OUTDIR/$(basename "$f") ($(du -h "$f" | cut -f1))"
  done
  TOTAL=$((TOTAL + ${#FILES[@]}))
done

echo "Done: $TOTAL file(s) written to $OUTDIR"