On Feb 14, 2009, at 7:25 PM, Wolfgang Schuster wrote:

Hi Thomas,

why don't you take a look at the OpenOffice export function, I saw it's possible to convert a document to xhtml and this could be a start for you.

Wolfgang

Hi Wolfgang, thanks for the suggestion! I had, in fact, tried the export functions (docbook and xhtml), but both drop too much formating: all italics etc. are silently dropped, and dynamical references are replaced with their values. So unless I can manage to hack the export xslt files, this doesn't seem possible. All best Thomas

Luigi and Khaled, thanks a lot for your replies! Luigi: I had a look at python lxml; it looks very powerful and interesting, and I will try and see if can make use of it. Why do you translate your xml sources into tex instead of using the mkiv mechanism for processing xml, is it because of speed? Khaled: I have to see if I can tweak the OpenOffice docbook converter to keep more of the formatting; in its default state, it drops too much important stuff... Right now, I have followed Patrick's advice. I've installed saxon9 and am writing a xslt stylesheet to translate the openoffice xml into a cleaner and easier to handle format. I'm making progress... Maybe we should put something like this on the wiki and make it a collaborative effort - I can only write rules for stuff that occurs in my documents, and that is of course only a subset of what OpenOffice has, so it would be good to add rules as people find interesting features. All best Thomas On Feb 15, 2009, at 10:39 AM, luigi scarso wrote:

If you know python http://wiki.services.openoffice.org/wiki/PyUNO_bridge http://opendocumentfellowship.com/projects/odfpy For xml the choice is http://codespeak.net/lxml/

A native xml db, with XQuery and python binding http://www.oracle.com/technology/products/berkeley-db/xml/index.html

And this is my experience : I'm programming in TeX (with context) , lua / python (they are similar) and xslt . For every project if I can I use lxml to manage xml sources, because it includes xslt but not viceversa. The goal is to translate xml in tex in the quickest way, and let mkiv to do the hard word. I have not a good feeling with xslt, because is not so powerful as lxml, and clearly is not a competitor of TeX .

If I need storage, dbxml is good, and XQuery+lxml is powerful enought .

OO has also docbook exporter http://www.docbook.org/ docbook is rich and with a good collection of xsl stylesheets to translate xml to html but maybe is ...too much .

-- luigi ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : https://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

On Feb 17, 2009, at 11:07 PM, luigi scarso wrote:

(sorry x my laziness) If I have a good xml , then mkiv is a good choice. As far I know, mkiv ~ xslt by lpeg, so "traditional" xml--( xslt )-->tex--( mkiv )-->pdf is like xml-->( mkiv )-->pdf Note that in the last chain one mixes xml+tex: if xml become complex, this can end in a messy situation.

Yes, you're right of course. I have a similar situation here: the xml produced by ooo is too messy, so I want to preprocess it to something that is easier to maintain and modify (e.g., I will, at some point, add index entries and a TOC); that's why I use xslt here. But I still produce xml which I process with mkiv.

But some documents need heavy preprocessing: for example, I have one that come from java classes serialization, and I need the power of python (lxml) to do a clean work . Also, if xml changes , I 've found that lxml is more flexible than xslt. In this case I have xml--( lxml )-->tex--( mkiv )-->pdf

The fact is that python and lua are not so differents, so I've to manage two languages (python+lua) and tex; with 'traditional' workflow you have to manage 3 languages xslt,lua and tex and subdivide responsability is not so easy as the former .

Interesting. I have tried to play around with python-lxml, but am having some problems to understand it. Just to give me an idea: how would you transform this: foo to this <emph>foo</emph> with lxml? lxml seems to object to the ":" in the tag, even though it's declared in the document. Thomas

On Thu, Feb 19, 2009 at 9:54 AM, Thomas A. Schmitz wrote:

On Feb 17, 2009, at 11:07 PM, luigi scarso wrote:

...

(sorry x my laziness) If I have a good xml , then mkiv is a good choice. As far I know, mkiv ~ xslt by lpeg, so "traditional" xml--( xslt )-->tex--( mkiv )-->pdf is like xml-->( mkiv )-->pdf Note that in the last chain one mixes xml+tex: if xml become complex, this can end in a messy situation.

Yes, you're right of course. I have a similar situation here: the xml produced by ooo is too messy, so I want to preprocess it to something that is easier to maintain and modify (e.g., I will, at some point, add index entries and a TOC); that's why I use xslt here. But I still produce xml which I process with mkiv.

...

But some documents need heavy preprocessing: for example, I have one that come from java classes serialization, and I need the power of python (lxml) to do a clean work . Also, if xml changes , I 've found that lxml is more flexible than xslt. In this case I have xml--( lxml )-->tex--( mkiv )-->pdf

The fact is that python and lua are not so differents, so I've to manage two languages (python+lua) and tex; with 'traditional' workflow you have to manage 3 languages xslt,lua and tex and subdivide responsability is not so easy as the former .

Interesting. I have tried to play around with python-lxml, but am having some problems to understand it. Just to give me an idea: how would you transform this:

foo

to this

<emph>foo</emph>

with lxml? lxml seems to object to the ":" in the tag, even though it's declared in the document.

Thomas

t.xml: <foo xmlns:text="urn:oasis:names:tc:opendocument:xmlns:text:1.0"> foo </foo> # python Python 2.5.2 (r252:60911, Jul 31 2008, 17:28:52) [GCC 4.2.3 (Ubuntu 4.2.3-2ubuntu7)] on linux2 Type "help", "copyright", "credits" or "license" for more information.

...

...

from lxml import etree tree = etree.parse(file('t.xml')) foo = tree.getroot() foo.tag 'foo'

[child.tag for child in foo.iterdescendants() ] ['{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span'] print foo.iterdescendants.__doc__ iterdescendants(self, tag=None)

Iterate over the descendants of this element in document order. As opposed to ``el.iter()``, this iterator does not yield the element itself. The generated elements can be restricted to a specific tag name with the 'tag' keyword.

...

...

FOO = etree.Element("FOO") emph = etree.Element("emph") [child.tag for child in foo.iterdescendants(tag = '{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span' ) ]

['{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span']

...

span = [child for child in foo.iterdescendants(tag = '{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span' ) ][0] emph.text = span.text FOO.append(emph) etree.tostring(FOO) '<FOO><emph>foo</emph></FOO>'

http://codespeak.net/lxml/tutorial.html http://codespeak.net/lxml/api.html -- luigi

On Thu, Feb 19, 2009 at 12:53 PM, Thomas A. Schmitz wrote:

On Feb 19, 2009, at 11:39 AM, luigi scarso wrote:

...

...

...

...

FOO = etree.Element("FOO") emph = etree.Element("emph") [child.tag for child in foo.iterdescendants(tag = '{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span' ) ]

['{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span']

...

...

...

span = [child for child in foo.iterdescendants(tag = '{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span' ) ][0] emph.text = span.text FOO.append(emph) etree.tostring(FOO)

'<FOO><emph>foo</emph></FOO>'

...

...

...

Excuse me for being dense: you mean all namespaces have to be explicitly expanded?

see http://codespeak.net/lxml/tutorial.html#namespaces -- luigi

On Fri, Feb 20, 2009 at 4:09 PM, Thomas A. Schmitz wrote:

On Feb 19, 2009, at 3:10 PM, luigi scarso wrote:

...

see http://codespeak.net/lxml/tutorial.html#namespaces

Luigi,

thanks so much for your patient replies. I have now begun to play with python's lxml. It offers a lot, maybe too much for a beginner. One advantage for my immediate needs that I see is that it offers the possibility to use Python's regular expressions and control structures, so this may make coding easier to maintain and adapt that in the rather clumsy xslt syntax; it may be a big help for the rather messy OpenOffice xml that I want to process.

also Python 2.5.2 (r252:60911, Jul 31 2008, 17:28:52) [GCC 4.2.3 (Ubuntu 4.2.3-2ubuntu7)] on linux2 Type "help", "copyright", "credits" or "license" for more information.

...

...

URI_OFFICE = "urn:oasis:names:tc:opendocument:xmlns:office:1.0" URI_STYLE = "urn:oasis:names:tc:opendocument:xmlns:style:1.0" URI_TEXT = "urn:oasis:names:tc:opendocument:xmlns:text:1.0" URI_TABLE = "urn:oasis:names:tc:opendocument:xmlns:table:1.0" URI_DRAW = "urn:oasis:names:tc:opendocument:xmlns:drawing:1.0" URI_FO = "urn:oasis:names:tc:opendocument:xmlns:xsl-fo-compatible:1.0" URI_XLINK = "http://www.w3.org/1999/xlink" URI_DC = "http://purl.org/dc/elements/1.1/" URI_META = "urn:oasis:names:tc:opendocument:xmlns:meta:1.0" URI_NUMBER = "urn:oasis:names:tc:opendocument:xmlns:datastyle:1.0" URI_PRESENTATION = "urn:oasis:names:tc:opendocument:xmlns:presentation:1.0" URI_SVG = "urn:oasis:names:tc:opendocument:xmlns:svg-compatible:1.0" URI_CHART = "urn:oasis:names:tc:opendocument:xmlns:chart:1.0" URI_DR3D = "urn:oasis:names:tc:opendocument:xmlns:dr3d:1.0" URI_MATH = "http://www.w3.org/1998/Math/MathML" URI_FORM = "urn:oasis:names:tc:opendocument:xmlns:form:1.0" URI_SCRIPT = "urn:oasis:names:tc:opendocument:xmlns:script:1.0" URI_OOO = "http://openoffice.org/2004/office" URI_OOOW = "http://openoffice.org/2004/writer" URI_OOOC = "http://openoffice.org/2004/calc" URI_DOM = "http://www.w3.org/2001/xml-events" URI_XFORMS = "http://www.w3.org/2002/xforms" URI_XSD = "http://www.w3.org/2001/XMLSchema" URI_XSI = "http://www.w3.org/2001/XMLSchema-instance" URI_FIELD = "urn:openoffice:names:experimental:ooxml-odf-interop:xmlns:field:1.0"

...

...

NSMAP_OO = { "office" : URI_OFFICE, "style" : URI_STYLE, "text" : URI_TEXT, "table" : URI_TABLE, "draw" : URI_DRAW, "fo" : URI_FO, "xlink" : URI_XLINK, "dc" : URI_DC, "meta" : URI_META, "number" : URI_NUMBER, "presentation" : URI_PRESENTATION, "svg" : URI_SVG, "chart" : URI_CHART, "dr3d" : URI_DR3D, "math" : URI_MATH, "form" : URI_FORM, "script" : URI_SCRIPT, "ooo" : URI_OOO, "ooow" : URI_OOOW, "oooc" : URI_OOOC, "dom" : URI_DOM, "xforms" : URI_XFORMS, "xsd" : URI_XSD, "xsi" : URI_XSI, "field" : URI_FIELD, }

...

...

from lxml import etree

...

...

tree = etree.parse(file('t.xml'))

...

...

...

...

foo = tree.getroot()

...

...

[child.tag for child in foo.iterdescendants(tag = '{%s}span'%URI_TEXT ) ] ['{urn:oasis:names:tc:opendocument:xmlns:text:1.0}span']

give a look at http://opendocumentfellowship.com/projects/odfpy too -- luigi