sxw2rml cannot support for Simplified Chinese Version OpenOffice 1.0 document

[mrshelly]

I used the following python script convert sxw(OpenOffice 1.0 document) to rml document.

<pre>

import zipfile,sys
from pyopenoffice import PyOpenOffice
import StringIO
from lxml import etree
import xml.dom.minidom

import libxslt
import libxml2

fname = r'c:\test.sxw'
xsl_file = './normalized_oo2rml.xsl'
z = zipfile.ZipFile(fname, 'r')
mimetype = z.read('mimetype')
if mimetype.split('/')[-1] == 'vnd.oasis.opendocument.text' :
    xsl_file = './normalized_odt2rml.xsl'

xsl = file(xsl_file).read()
tool = PyOpenOffice('.', save_pict = False)
sxw_file = fname
res = tool.unpackNormalize(sxw_file)

styledoc = libxml2.parseDoc(xsl)

style = libxslt.parseStylesheetDoc(styledoc)
doc = libxml2.parseMemory(res,len(res))
result = style.applyStylesheet(doc, None)
print result

</pre>

There are some bug of minidom python extended, and I fixed it.

@tiny_sxw2rml.pdf (5.x) or @openerp_sxw2rml.pdf I found the code...

<pre>
        styles_styles = self.styles_dom.getElementsByTagName("style:style")

</pre>

I fixed it like :

<pre>
        ....
        styles_styles = []
        styles_styles = styles_styles + self.styles_dom.getElementsByTagName("style:style")
        styles_styles = styles_styles + self.styles_dom.getElementsByTagName("style:font-decl")
        ....
</pre>

and some trouble with "content_styles" variable...

@normalized_oo2rml.xsl document. I found the code:

<pre>
<xsl:when test="not($fontName='') and boolean($fontName)">

....
    <xsl:when test="contains($fontName,'Courier')">

...
    <xsl:when test="contains($fontName,'Helvetica') or contains($fontName,'Arial') or contains($fontName,'Sans')">

...
    <xsl:otherwise> <-------------------- Otherwise 1

...
<xsl:otherwise> <-------------------- Otherwise 2
...
</pre>

In Simplified Chinese Version OpenOffice 1.0 document, The "fontName" is "宋体", "黑体".
I found in my "test.sxw" file, the normalized_oo2rml.xsl match the "Otherwise 2", the sxw file's "宋体" fontName be replaced with "Times-Roman"..

Then, How to fixed it and add docini/registerFont node to generated rml file. order to let OpenERP to support the Simplified Chinese Version OpenOffice 1.0 document can be convert to rml file.

Thanks...

mrshelly
2011/05/25