Problem in XSLT...

Jon2078 · December 10th, 2011, 08:35 AM

Hi All,

I have a requirment to impelment CockeâYoungerâKasami (CYK) algorithm in the XSLT and transform the XML to produce new XML file. Could you please tell me if is it possible in the XSLT.

If yes, could you please give me an idea or example code.

I have seen the implementaion of the this algorithm in 'Java' but not able to fine the example code in XSLT.

I have tried my self but not able to achieve what I am after using XSLT.

Many thanks in advance. Your help much appriciated.

Regards
John

samjudson · December 10th, 2011, 09:05 AM

Seeing as it is unlikely anyone even knows what the CYK algorithm is it is unlikely we will be able to help you with such a request unless you give us more to go on.

Generally you specify the input XML, what you would like the resulting output XML to look like and what attempts you have made to perform the transformation yourself.

Jon2078 · December 10th, 2011, 09:59 AM

Hi Sam,

Thank you for your quick reply.

To have a brief idea about CYK algorithm please have a look on the link below:

http://en.wikipedia.org/wiki/CYK_algorithm

There are two input xml like below I have to pass sentance.xml in the xslt and then based on the words in each sentance I have to read the values at runtime from the Rule.xml file and then produce the new XML given below.

1) sentance.xml

Code:

<?xml version="1.0" encoding="UTF-8"?>
<sentances>
    <sentance>dog bark</sentance>
    <sentance>cat drink milk</sentance>
</sentances>

2) Rule.xml

Code:

<?xml version="1.0" encoding="UTF-8"?>
<allrules>
    <rules>
        <rule cat="s">
            <rulechild cat="np"/>
            <rulechild cat="vp"/>
        </rule>
        <rule cat="vp">
            <rulechild cat="vt"/>
            <rulechild cat="np"/>
        </rule>
        <rule cat="vp">
            <rulechild cat="vi"/>
        </rule>	
    </rules>
    <words>
        <word cat="vi">bark</word>
        <word cat="vt">drink</word>
        <word cat="pn">dog</word>
        <word cat="pn">cat</word>
        <word cat="pn">milk</word>
    </words>
</allrules>

OutPut XML should be like below:

Code:

<trees>
    <tree>
        <sentace>dog bark</sentace>
        <node cat="s">
            <node cat="np">
                <word cat="pn">dog</word>
            </node>
            <node cat="vp">
                <word cat="vi">bark</word>
            </node>
        </node>
    </tree>
    <tree>
        <sentace>cat drink milk</sentace>
        <node cat="s">
            <node cat="np">
                <word cat="pn">cat</word>
            </node>
            <node cat="vp">
                <word cat="vt">drink</word>
                <node cat="np">
                    <word cat="pn">milk</word>
                </node>
            </node>
        </node>
    </tree>
</trees>

Could it be possible to implement the CYK algorithm and produce the above out using XSLT.

Any idea or example code will be very greateful.

Thanks again.
John

Jon2078 · December 10th, 2011, 03:37 PM

Someone please shed some light on this. really appriciated.

Thanks
John

mhkay · December 10th, 2011, 07:49 PM

Reverse engineering a functional, declarative program from a procedural program is hard. You can simulate the algorithm, but it may be very inefficient. Very often it's better to start with the problem statement (which in this case seems to be designing a parser for a class of grammars) and to choose or design an algorithm for solving that problem which is better suited to a functional approach.