(C) 2012 by Damir Cavar and Malgosia Cavar
Project Gutenberg books to TEI XML conversion
This repository contains code for the automatic conversion of Project Gutenberg books to the TEI XML format. We focus on the generation of valid TEI Lite P5 XML from HTML-sources. The code is written in Java. It might use some Java 7 specific elements and constructions, but should be easily adaptable to Java 6 or earlier versions.
You might need the following components to convert the Project Gutenberg files yourself:
- Java SE 7
- The TEI Subversion repository on SourceForge
- Some document conversion tool
For the conversion of ODT-documents to TEI XML we make use of XSLT-scripts that are part of the TEI@Sourceforge package. You can check out a local copy of the trunk using the following command:
svn co https://tei.svn.sourceforge.net/svnroot/tei/trunk ./TEI
For further details, see the links to the repository here:
The necessary components will be in the Stylesheets subfolder. In particular relevant is the odttotei script. You might have to change certain paths in the script, or provide appropriate command line parameters when invoking it.
In addition to the XSLT-scripts in the TEI folder you will most likely need Saxon. If you use oXygen, you should provide the path to the oXygen-lib-folder via command line to odttotei or directly in your adjusted script (e.g. a version of odttotei).
We make use of the textutil tool that is distributed with the recent versions of Mac OS X. textutil makes batch conversion of different document types easy. We use textutil to convert the Project Gutenberg HTML-files to ODT.
You might want to try alternative conversion strategies, for example using:
- pandoc, a universal document converter that is available for all major platforms.
- OpenOffice or LibreOffice via command line for batch processing. You will find a lot of descriptions of the command line usage online, see for example here...
Since the PG2TEI-code is a quick and dirty implementation of the conversion pipeline, with a very defensive coding strategy, avoiding complications that might improve the stability, but would cost coding time, there are some things in the code that need specific adaptation. You might experience crashes and error messages for individual files. We cannot avoid that. The conversion runs quite stable, and restarting the converter skips already available target files.
Follow the instructions in these documents to set up the conversion process for your specific environment:
If you discover serious bugs or problems with the code, please send us a message. Thanks!
The code is made available under the Apache 2.0 License as is. See LICENSE.md.