need a little more information in the GraphML

[Tara]

There are a couple of things I would like in the GraphML output which would be difficult to calculate myself, and which I think CollateX already calculates anyway.

At the moment, the GraphML uncouples transpositions, to keep the graph unidirectional. This is fine, but I would like an added key in the duplicated nodes, to indicate that they actually did get matched together. e.g. 'march of drought' vs. 'drought of march':

...
    <key attr.name="number" attr.type="int" for="node" id="d1"/>
    <key attr.name="token" attr.type="string" for="node" id="d0"/>
+ <key attr.name="identity" attr.type="int" for="node" id="d2"/>
...
        <node id="14">
            <data key="d0">drought</data>
            <data key="d1">14</data>
+ <data key="d2">18</data>
        </node>
        <node id="15">
            <data key="d0">march</data>
            <data key="d1">15</data>
+ <data key="d2">17</data>
        </node>
        <node id="16">
            <data key="d0">of</data>
            <data key="d1">16</data>
        </node>
        <node id="17">
            <data key="d0">march</data>
            <data key="d1">17</data>
+ <data key="d2">15</data>
        </node>
        <node id="18">
            <data key="d0">drought</data>
            <data key="d1">18</data>
+ <data key="d2">14</data>
        </node>
...

Second, it is easy to see visually which nodes belong in 'columns' together; I know CollateX can put the graph into an alignment table; so I wonder if this alignment information can get put into the GraphML? A fourth data key called 'column' or some such, with a numeric ID for the columns, would do the trick.

[Tara (tla)]

Fixed in branch ~tla/collatex/graphml (off of 1.0release branch)