Ubuntu – How to convert an ODT file to a PDF

A quick and dirty example using code from the pyinotify project [http://github.com/seb-m/pyinotify]

You will need to change the WATCHED_DIR to your directory containing ODT files. Also remember to install unoconv first.

# Notifier example from tutorial
#
# See: http://github.com/seb-m/pyinotify/wiki/Tutorial
#
# odtwatcher.py

import os
import pyinotify
import subprocess

WATCHED_DIR = '/tmp/test'

wm = pyinotify.WatchManager()  # Watch Manager
mask = pyinotify.IN_MODIFY

class EventHandler(pyinotify.ProcessEvent):
    def process_IN_MODIFY(self, event):
        fname = event.pathname            
        if os.path.splitext(fname)[1] == '.odt':
            print 'MODIFIED: ', fname
            args = ['unoconv', '-f', 'pdf', fname]
            try:
                subprocess.Popen(args)
            except OSError as e:
                print 'Could not convert file %s to PDF. Error %s' % (fname, str(e))

if __name__ == '__main__':
    handler = EventHandler()
    notifier = pyinotify.Notifier(wm, handler)
    wdd = wm.add_watch(WATCHED_DIR, mask, rec=True)        
    notifier.loop()

Save this as odtwatcher.py and then run it in the background

python odtwatcher.py &

Believe it or not, with complex documents and lots of packages included, I got much better results with LaTeX2HTML than with LaTeX2RTF, Pandoc or TeX4ht.

latex2html texfile.tex -split 0 -no_navigation -info "" -address "" -html_version 4.0,unicode

This will generate a folder with the same texfile name, so you'll be able to convert the generated HTML to ODT:

libreoffice --headless --convert-to odt:"OpenDocument Text Flat XML" texfile/index.html

This will produce a index.odt file. Take a look at this answer to check how to use LibreOffice's convert filters.

Edit from comment discussion:

Although the method above works, it is very disappointing that the only way I found to generate a trully reliable document is using the PDF output from LaTeX on Adobe Acrobat Pro.