Johan van der Knijff bitsgalore

Background

An EPUB is just a ZIP container, but using a ZIP tool directly on a directory with content documents won't usually result in a valid EPUB. This is because the standard requires that:

The mimetype resource must appear as the first file in the container
The mimetype resource must be uncompressed

So to meet these requirements we must ZIP the files in a special way. This gist describes how to do this with InfoZip (which is the default ZIP tool on most Linux systems).

Let's suppose all content files are in a directory called /home/johan/epubPolicyTests/content/epub20_minimal/.

All snapshots archived 13 March 2015.

http://web.archive.org/web/20150312225312/http://production.openplanetsfoundation.org/blogs/2014-10-29-quattro-pro-dos-obsolete-format-last

http://web.archive.org/web/20150312225424/http://production.openplanetsfoundation.org/blogs/2014-10-23-running-archived-android-apps-pc-first-impressions

http://web.archive.org/web/20150312225455/http://production.openplanetsfoundation.org/blogs/2014-09-26-six-ways-decode-lossy-jp2

http://web.archive.org/web/20150312225523/http://production.openplanetsfoundation.org/blogs/2014-08-27-when-not-migrate-pdf-pdfa

Experimental attempt at getting organized ...

07/11/2024

HTML Text fragments

https://developer.mozilla.org/en-US/docs/Web/URI/Fragment/Text_fragments

Extension	Number of files in e-Depot	ID(s) Tika	Remarks
gif	34499095	- image/gif	GIF image
xml	12913388	- application/xml	XML (mostly metadata)
jpg	8197415	N/A^*	JPEG image
sml	7744829	- image/gif	GIF image with unusual extension
pdf	7577414	- application/pdf	PDF
raw	2045662	- text/plain	Text file
tif	715509	- image/tiff	TIFF image
oa3	296101	- text/plain	Looks like SGML (oases, Kluwer). See also: Publisher Data Formats. Metadata.

Extension	ID Tika	Remarks
cif	text/plain	Crystallographic Information File
csv	text/csv	Comma-separated values
mol	text/plain	MDL Molfile
tdb	text/plain	Thermo-Calc Database Format
r	text/x-rsrc	R source code
m	text/x-objcsrc	Objective-C source code

Category	Rendering software in reading rooms	Formats accessible in reading rooms?
Image formats	MS Paint, Windows Photoviewer	Yes
PDF	Adobe Acrobat	Yes
Web formats	Internet Explorer, Google Chrome	Yes
Office formats	Microsoft Office	Yes (support for old Office formats presently not clear)
Audio	Windows Media Player, VLC Media Player	No (hardware in reading rooms doesn't support audio)
Video	Windows Media Player, VLC Media Player	Partially (hardware in reading rooms doesn't support audio)
Metadata	Internet Explorer, Notepad, Wordpad	Yes
Executables, installers, system files	Not applicable	No

	<?xml version="1.0"?>
	<!--
	Schematron jpylyzer schema: verify if JP2 conforms to
	KB's profile for access copies (A.K.A. KB_ACCESS_LOSSY_01/07/2014)
	Simplified version for Geheugen Van Nederland migration, omits specific requirements for
	resolution, colour space,compression ratio, XML box and codestream comment.
	Johan van der Knijff, KB / National Library of the Netherlands , 18 March 2014.

	-->
	<s:schema xmlns:s="http://purl.oclc.org/dsdl/schematron">

	#!/bin/bash
	# Compute compression ratio of each JP2 in directory tree, report results to CSV file
	# Requires:
	# - jpylyzer
	# - xmllint (part of libxml library)
	#
	# If you're using Windows you can run this shell script within a Cygwin terminal: http://www.cygwin.com/
	#

	# Installation directory

	#!/bin/bash
	# Generate derived JP2s that discard user-specified number of quality layers from source JP2s
	# Requires:
	# - j2kdriver (Aware)
	#
	# If you're using Windows you can run this shell script within a Cygwin terminal: http://www.cygwin.com/
	#

	# Installation directory
	instDir="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"

	#!/gnuplot
	#
	# Histogram of compression ratio distribution
	#
	set terminal postscript enhanced landscape
	set output "histogram.ps"
	set size ratio 0.5
	set key top right
	set xlabel "Compression ratio"
	set ylabel "Number of files"