xmParse - A C++ XML-to-MATLAB Parser
Adnan Duric
aduric@gmail.com
====================================

A Basic XML Parser that takes as input an XML file and creates posting entries for each node, which is converted to MATLAB sparse matrices.

It generates tables based on the relationships between parent-child nodes with recurrance information. This in turn produces sparse matrices that can be loaded into a MATLAB environment.

Example
-------

An XML file:

<DOCS>
<DOC>
<ENT name='entity1' value='val1'><SUBENT>something here</SUBENT></ENT>
<ENT name='entity2' value='val2'><SUBENT>something there</SUBENT><SUBENT>something there</SUBENT></ENT>
</DOC>
<DOC>
<ENT name='entity4' value='val4'><SUBENT>something everywhere</SUBENT></ENT>
</DOC>
</DOCS>

- Since it is clear from the structure fo the XML document above that there is a 3-deep hierarchy, we desire the data to be represented in the following MATLAB matrices:

vocabulary (an array representing all the words)

| 0 | 'entity1'    |
| 1 | 'val1'       |
| 2 | 'something'  |
| 3 | 'here'       |
| 4 | 'entity2'    |
| 5 | 'val2'       |
| 6 | 'there'      |
| 7 | 'entity3'    |
| 8 | 'val3'       |
| 9 | 'everywhere' |


subent (a SUBENT x VOCAB matrix)

| 0 | 0 | 1 | 1 | ... size of vocabulary
| 0 | 0 | 1 | 0 | 1 | 0 | 1 | ...
| 0 | 0 | ...
| 0 | 0 | 1 | 0 | ...
... number of SUBENT nodes


ent_name (ENT x vocabulary sparse matrix where data represents all entity names)

| 1 | 0 | 0 | ... size of vocabulary
| 0 | 0 | 0 | 0 | 1 | 0 | ...
| 0 |
... number of ENT nodes


ent_val (ENT x vocabulary sparse matrix where data represents all entity values)

| 0 | 1 | 0 | ... size of vocabulary
| 0 | 0 | 0 | 0 | 0 | 1 | ...
| 0 |
... number of ENT nodes



ent_subent (ENT x SUBENT matrix)

| 1 | 0 | 0 | 0 |
| 0 | 1 | 1 | 0 |
| 0 | 0 | 0 | 1 |


doc_ent (DOC x ENT sparse matrix where data is the occurance count)

| 1 | 1 | 0 |
| 0 | 0 | 1 |


These matrices can then be used to manipulate the count data as you see fit in the MATLAB environment.