#!/bin/sh -
# $Id: webmgp2pdf,v 1.2 2004/02/17 22:27:53 sra Exp $
#
# Hack to generate PDF files from a MagicPoint presentation that's
# been posted on a web site.
#
# The basic idea here is to snarf the web version of a MagicPoint
# presentation using wget, extract the screen images, then combine the
# images into a PDF file.  The resulting PDF is huge, because it's all
# raster graphics, but it's easier to ship around than a directory
# full of html and image files would be.
#
# This uses the NetPBM toolkit, and GhostScript.  It's a kludge.  Feel
# free to do better, then send me what you did so I can use it too.
#
# Portions of this (eg, the anytopnm usage) really need rewriting.
#
# This program is hereby explictly placed in the public domain as
# Beer-Ware.  If we meet some day and you think this program is worth
# it, you can buy me a beer.  Your mileage may vary.  We decline
# responsibilities, all shapes, all sizes, all colors.  If this
# program breaks, you get to keep both pieces.

# Filenames
usage="usage: $0 url output.pdf"
url="${1?"${usage}"}"
output="${2?"${usage}"}"

# No user servicable parts past this point.

# Always mount a scratch monkey
tmp="tmp.$$"
trap "rm -rf '$tmp'" 0 1 2 3 13 15
mkdir "$tmp"

# Snarf the files into our temp directory
wget --mirror --relative --no-parent --no-directories --directory-prefix="$tmp" "$url" || {
  echo 1>&2 "$0: wget failed, giving up"
  exit 1
}

# We now have a temporary directory containing PNG or JPG renderings
# of our slides, along with a bunch of stuff we don't care about.

# Find the image files
find "$tmp" -type f \( -name 'mgp*.png' -o -name 'mgp*.jpg' \) ! -name '*.idx.*' >"$tmp/files" || {
  echo 1>&2 "$0: couldn't find images in downloaded presentation, giving up"
  exit 1
}

# Extract and check the image geometry
while read i; do bash anytopnm 2>/dev/null "$i" | pnmfile; done <"$tmp/files" | awk '
  NR == 1 { x = $4; y = $6 }
  $4 != x || $6 != y { exit 1 }
  END { print x "0x" y "0" }
' >"$tmp/geometry" || {
  echo 1>&2 "$0: downloaded images weren't what i expected, giving up"
  exit 1
}

# Convert the images to Postscript, then to PDF.  It'd be nice to
# include the thumbnails, but I don't know any easy way to do that.

while read i ; do bash anytopnm "$i" | pnmtops -nocenter -noturn -dpi 72 -equalpixels; done <"$tmp/files" |
ps2pdf13 "-g$(cat "$tmp/geometry")" - "$tmp/output.pdf"

# That's it.
mv -f "$tmp/output.pdf" "$output"