# Need to be imported first to avoid dll loading problems.
import pywintypes
import pythoncom

import os.path

import win32com, sys, string, win32api, traceback, re, tempfile, os
import win32com.client
from win32com.client import gencache
from win32com.client import constants, Dispatch

from Products.PortalTransforms.libtransforms.commandtransform import commandtransform
from Products.PortalTransforms.libtransforms.utils import bodyfinder, scrubHTML

class document(commandtransform):

    def __init__(self, name, data):
        """Initialization: create tmp work
        directory and copy the document into a file"""
        commandtransform.__init__(self, name)
        name = self.name()
        if not name.endswith('.doc'):
            name = name + ".doc"
        self.tmpdir, self.fullname = self.initialize_tmpdir(data, filename=name)

    def convert(self):
        try:
            # initialize COM for multi-threading, ignoring any errors
            # when someone else has already initialized differently.
            pythoncom.CoInitializeEx(pythoncom.COINIT_MULTITHREADED)
        except pythoncom.com_error:
            pass

        word = Dispatch("Word.Application")
        word.Visible = 0
        word.DisplayAlerts = 0
        doc = word.Documents.Open(self.fullname)
        # Let's set up some html saving options for this document
        doc.WebOptions.RelyOnCSS = 1
        doc.WebOptions.OptimizeForBrowser = 1
        doc.WebOptions.BrowserLevel = 0 # constants.wdBrowserLevelV4
        doc.WebOptions.OrganizeInFolder = 0
        doc.WebOptions.UseLongFileNames = 1
        doc.WebOptions.RelyOnVML = 0
        doc.WebOptions.AllowPNG = 1
        # And then save the document into HTML
        doc.SaveAs(FileName = "%s.htm" % (self.fullname),
                   FileFormat = 8) # constants.wdFormatHTML)

        # TODO -- Extract Metadata (author, title, keywords) so we
        # can populate the dublin core
        # Converter will need to be extended to return a dict of
        # possible MD fields

        doc.Close()
        # word.Quit()

    def html(self):
        htmlfile = open(self.fullname + '.htm', 'r')
        html = htmlfile.read()
        htmlfile.close()
        html = scrubHTML(html)
        body = bodyfinder(html)
        return body

## This function has to be done. It's more difficult to delete the temp
## directory under Windows, because there is sometimes a directory in it.
##    def cleanDir(self, tmpdir):