What's inside an EPUB file

Electronic books or "ebooks" are a great way to share technical documentation and other written materials in a portable format. A few ebook file formats have come and gone, but the most popular ebook format remains the EPUB, short for "electronic publication." EPUB files are supported across a variety of systems and software, and are effectively the de facto standard for ebook publication today.

The EPUB file format is an open standard based on XHTML for content and XML for metadata, contained in a zip file archive. And because everything is based on open standards, we can use common tools to create or examine EPUB files. Let's explore an EPUB file to learn more about it. For this example, I'll use a copy of my ebook Tips and Tricks for C Programmers, published by Opensource.com last year.

To understand how an EPUB file is put together, follow the process flow of an ebook reader:

1. Verify it is an EPUB
2. Locate the content
3. Identify the order of the content

1. Verify it is an EPUB

Because EPUB files are XHTML content and XML metadata in a zip file, you can start with the unzip command to examine the EPUB from the command line:

$ unzip -l osdc_Jim-Hall_C-Programming-Tips.epub 
Archive:  osdc_Jim-Hall_C-Programming-Tips.epub
 Length      Date    Time    Name
---------  ---------- -----   ----
      20  06-23-2022 00:20   mimetype
    8259  06-23-2022 00:20   OEBPS/styles/stylesheet.css
    1659  06-23-2022 00:20   OEBPS/toc.xhtml
    4460  06-23-2022 00:20   OEBPS/content.opf
   44157  06-23-2022 00:20   OEBPS/sections/section0018.xhtml
    1242  06-23-2022 00:20   OEBPS/sections/section0002.xhtml
   22429  06-23-2022 00:20   OEBPS/sections/section0008.xhtml
[...]
    9628  06-23-2022 00:20   OEBPS/sections/section0016.xhtml
     748  06-23-2022 00:20   OEBPS/sections/section0001.xhtml
    3370  06-23-2022 00:20   OEBPS/toc.ncx
    8308  06-23-2022 00:21   OEBPS/images/image0011.png
    6598  06-23-2022 00:21   OEBPS/images/image0009.png
[...]
   14492  06-23-2022 00:21   OEBPS/images/image0005.png
     239  06-23-2022 00:20   META-INF/container.xml
---------                     -------
  959201                     41 files

This EPUB contains a lot of files, but much of this is content. Ebook readers start with the mimetype file to verify that the ebook really is an EPUB. The mimetype file describes the MIME type of the ebook (an EPUB file inside a zip archive) and is stored as an uncompressed file at the start of the zip archive:

application/epub+zip

2. Locate the content

To locate the content, ebook readers then look for the META-INF/container.xml file. This is a brief XML document that indicates where to find the content. For this EPUB file, the container.xml file looks like this:

<?xml version="1.0" encoding="UTF-8"?>
 <container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
   <rootfiles>
     <rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/>
 </rootfiles>
</container>

To make the container.xml file easier to read, I split the single line into multiple lines and added some spacing to indent each line. XML files don't really care about extra white space like new lines and spaces, so this extra spacing doesn't affect the XML file.

The container.xml file says the root of the EPUB starts with the content.opf file in the OEBPS directory. Many ebooks use an OEBPS directory, but they can locate the file anywhere. The "OEBPS" name is an abbreviation of the Open Electronic Book Publication Structure. The content.opf file is really just another XML file; it uses the OPF extension is because EPUB is based on the Open Packaging Format.

3. Identify the order of the content

The content.opf file contains a complete manifest of the EPUB contents, plus an ordered table of contents, with references to find each chapter or section. The content.opf file for this EPUB is quite long, so I'll show just a short section of it as an example.

The XML data is contained within a <package> block, which itself has a <metadata> block, the <manifest> data, and a <spine> block that contains the ebook's table of contents:

<?xml version="1.0" encoding="UTF-8"?>
<package unique-identifier="unique-identifier" version="3.0" xmlns="http://www.idpf.org/2007/opf" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:opf="http://www.idpf.org/2007/opf">
<metadata>
<dc:identifier id="unique-identifier">osdc002</dc:identifier>
<dc:title>Tips and Tricks for C Programming</dc:title>
<dc:creator>Jim Hall</dc:creator>
<dc:language>English</dc:language>
<meta property="dcterms:modified">2022-06-23T12:09:13Z</meta>
<meta content="LibreOffice/7.3.0.3$Linux_X86_64 LibreOffice_project/0f246aa12d0eee4a0f7adcefbf7c878fc2238db3 (libepubgen/0.1.1)" name="generator"/>
</metadata>
<manifest>
...
<item href="https://opensource.com/sections/section0001.xhtml" id="section0001" media-type="application/xhtml+xml"/>
<item href="https://opensource.com/images/image0003.png" id="image0003" media-type="image/png"/>
<item href="https://opensource.com/styles/stylesheet.css" id="stylesheet.css" media-type="text/css"/>
<item href="https://opensource.com/toc.ncx" id="toc.ncx" media-type="application/x-dtbncx+xml"/>
...
</manifest>
<spine toc="toc.ncx">
<itemref idref="section0001"/>
<itemref idref="section0002"/>
<itemref idref="section0003"/>
...
</spine>
</package>

You can see in the <metadata> block that this EPUB was created with LibreOffice version 7.3.0.3 on Linux, which can export content as EPUB files.

You can match up the data to see where to find each section. That’s how EPUB readers do it. For example, the first item in the table of contents references section0001 which is defined in the <manifest> as located in the sections/section0001.xhtml file. The file doesn’t need to be named the same as the idref entry, but that’s how LibreOffice Writer’s automated process created the file. 

EPUB files are a great way to publish content using an open format. The EPUB file format is XML metadata with XHTML content, inside a zip container. Most technical writers use tools to create EPUB files. But because EPUB is based on open standards, you can create your own EPUB files or pull it apart to see what's inside.