A simple-minded Jython class to aid hand-converting a collection of HTML files into a valid EPUB document.

Some virtual methods provided, to allow customization of generated content.

The tool requires Adobe epubcheck (see example below).

• A set of HTML files
• Write a short script (around 100 lines) to subclass the EPUBTool with a content processor and manifest generator. You can optionally provide a cover page and a ToC JSON file of the format:

[[html_file, nesting_level, entry_text], ... ]

Or, alternatively, generate one from a CHM *.hhc file using the script provided.

• For the exquisite results, you can also write an image file filter, pack processor or even invoke HTML to XHTML content convertor!
• NB: you could individually suppress specific validation messages, or even silence them altogether (see the launcher script).
• Run the launcher script.
• Voilla, you've got the EPUB book!

Would recommend to run it from under the NailGun Java framework to greatly speed up iterative tasks.

The module is intended to be imported from some other Jython script as a class. See the module file for the class details and usage example.

For example, the NailGun startup script could contain the following:

...
EPUBCHECK=some_root/text-tools/epubcheck-4.0.2
CLASSPATH="${EPUBCHECK}/lib/*:${EPUBCHECK}/*:${CLASSPATH}"
export CLASSPATH
nohup java com.martiansoftware.nailgun.NGServer &
sleep 2
# Set up aliases
ng ng-alias ng-epubcheck com.adobe.epubcheck.tool.Checker
ng ng-alias ng-jython org.python.util.jython
...

You can then run the tool as follows:

ng-jython epubtool.py example example.epub