The regexpchainer is there for this purpose, it take in a configuration file " /.dashboard/backend-data/regexchainer/config.xml" a list of regular expression and what should be done with them.
You can specify chain in two ways:
<Chain Type="url" Regexp="http://www.google.com/search\?&q=(.*)">
that will chain all clue from URL type to KEYWORD. extracting anything after the "q=" in a google query.
<Chain From="textblock" To="measure" Regex="(\b\d+((.|,)\d+)? ?oF)\b" Ignore="2,3"/>
will extract Fahrenheit MEASURE from TEXBLOCK.
if no Type attribute is specified, the parser expect the From and To tags; if Type is specified, KEYWORD clues will be generated. The regexp is in Perl5 format (the one supported by mono). Each group (between parentheses) will generate a new clue, except the one specified in the Ignore attribute. (Note that there is no group 0)
everything between quote should be escaped (& => &)
(actually the conversion backend relies on this chainer to extract measure)
