Re: [docbook-apps] Problem with tables and roundtrip to WordML

[David Hinds <dhinds@23andme.com>]

Incidentally, I did get tables working fairly well without too much work, though I can't vouch much for how I did it, due to my limited XSL knowledge.

In dbk2wp.xsl, my changes were:

- implement an 'align' parameter for <para> and <simpara> elements to be passed to make-paragraph
- changed the <entry> template to recognize the align= attribute and pass it through the call to apply-templates for the cell contents



- changed the <table> template to pass a role= attribute to make-table as the 'style' parameter

In dbk2wordml.xsl, my changes were:
- in make-paragraph, use 'align' to generate a paragraph alignment element



- in make-table, use 'style' parameter to select a table style

As straight bug fixes, there are several spots where <simpara> elements are not handled properly (i.e., in <entry> and <listitem>), so I made trivial fixes to those templates.� There also seem to be a couple spots in dbk2wp.xsl where apply-templates is missing a mode='doc.body' attribute (in the templates for admonitions and for varlistentry).� Without that, I get errors like "simpara encountered in note, but no template matches", or "varlistentry encountered in variablelist, but no template matches".� 



I'm actually working with stylesheets that were modified to generate Word 2007 OOXML, that I found here:

http://www.brians-brain.org/projects/org-to-docx.html


There seems to be little development activity on these stylesheets, so I'm not sure how to proceed from here.� Some of my changes would apply easily to the official distribution, though I have not tried to implement the reverse transformations since I don't have a need for those.� Are patches welcome?

-- Dave

On Mon, Feb 28, 2011 at 12:56 PM, Steve Ball <Steve.Ball@explain.com.au> wrote:

Hi David,

Although I have not yet checked, I believe the problem is that the content of the cells is not in a <para>. That is, your DocBook should look like:

<thead><row><entry><para>a1</para></entry><entry><para>a2</para></entry></row></thead>

��� <tbody><row><entry><para>b1</para></entry><entry><para>b2</para></entry></row></tbody>

As for being able to produce professional looking tables, there is some work to do in setting up table column widths and borders properly. However, YMMV.

Cheers,

Steve Ball

On 01/03/2011, at 7:16 AM, David Hinds wrote:

I'm trying to convert a very simple docbook file to WordML using the roundtrip XSL stylesheets.� I'm having trouble with tables: all cell contents are missing after the conversion.

Here is my input document:

<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN" "http://www.oasis-open.org/doc\
book/xml/4.5/docbookx.dtd">
<article lang="en">
<table frame='all'><title>Sample Table</title>
� <tgroup cols="2">
��� <colspec colname='c1'/><colspec colname='c2'/>
��� <thead><row><entry>a1</entry><entry>a2</entry></row></thead>
��� <tbody><row><entry>b1</entry><entry>b2</entry></row></tbody>
� </tgroup>
</table>
</article>

This renders ok in HTML with:

xsltproc --path .../docbook-xsl/html -o test.html \
� docbook.xsl test.xml

but with this:

xsltproc --path .../docbook-xsl/roundtrip -o test.wml \
� --stringparam wordml.template template.xml \
� dbk2wordml.xsl test2.xml

the cells in the resulting table are all empty, i.e. the resulting WordML looks like:

����������� <w:tc>
������������� <w:tcPr>
��������������� <w:tcW w:w="" w:type="dxa"/>
������������� </w:tcPr>
������������� <w:p/>
����������� </w:tc>

I get the same result with either xsltproc or saxon.� I am using docbook-xsl-1.76.1.

My long term goal is to be able to programmatically generate reports as templated, styled Word documents from asciidoc source files, with docbook as an intermediate format, in a linux environment.� I need relatively few features: headers, body text, simple bullet lists, PNG images, and reasonably professional-looking tables (where that means I need control over column alignment and borders).�

-- Dave