#!/usr/bin/env python
"""
A simple little script to convert Microsoft Word files into Konqueror .war
archives (an HTML file and it's dependencies inside a renamed .tar file)
using wvHtml.

TODO:
- Add a file magic check to identify and rename RTF files with .doc extensions.
"""

import os, shutil, sys, tarfile, tempfile

def do_conversion(fldrname, target):
  """Put the contents of a folder into a war archive."""
  warfile = tarfile.open(target, 'w')
  for filename in os.listdir(fldrname):
	print "--- Adding " + filename
  	warfile.add(os.path.join(fldrname,filename), filename)
  warfile.close()

if __name__ == "__main__" and len(sys.argv) > 1:
	for source in sys.argv[1:]:
		filepath = os.path.abspath(source)
		filename = os.path.split(filepath)[1]
		dirname  = tempfile.mkdtemp(prefix="doc2war-")
		htmlpath = os.path.join(dirname,'index.html')
		warbase  = os.path.splitext(filepath)[0]

		print "Reading " + filename
		if not os.spawnlp(os.P_WAIT,'wvHtml','wvHtml',filepath,htmlpath):
			flist = os.listdir(dirname)

			if len(flist) == 1 and os.stat(htmlpath).st_size:
				print "-- Creating %s.html" % os.path.split(warbase)[1]
				shutil.move(htmlpath, warbase + ".html")
			elif len(flist) > 1 and os.stat(htmlpath).st_size:
				print "-- Creating %s.war" % os.path.split(warbase)[1]
				do_conversion(dirname, warbase + ".war")
			else:
				print "!! Could not convert " + filename
				print "wvHtml was unable to read the given file"

		shutil.rmtree(dirname)
		if [x for x in ('.war', '.html') if os.path.exists(warbase + x)]:
			print "-- Deleting " + filename
			os.remove(filepath)
elif __name__ == '__main__':
	print "Usage: %s [input file] ..." % sys.argv[0]