package dr.evomodelxml.indel;
import dr.evomodel.indel.IstvanOperator;
import dr.evomodel.indel.TKF91Likelihood;
import dr.xml.*;
/**
*
*/
public class IstvanOperatorParser extends AbstractXMLObjectParser {
public static final String ALIGMENT_CHUNK_OPERATOR = "alignmentChunkOperator";
public String getParserName() {
return ALIGMENT_CHUNK_OPERATOR;
}
public Object parseXMLObject(XMLObject xo) throws XMLParseException {
TKF91Likelihood likelihood = (TKF91Likelihood) xo.getChild(TKF91Likelihood.class);
double weight = xo.getDoubleAttribute("weight");
double iP = xo.getDoubleAttribute("iP");
double exponent = xo.getDoubleAttribute("exponent");
double gapPenalty = xo.getDoubleAttribute("gapPenalty");
return new IstvanOperator(iP, exponent, gapPenalty, weight, likelihood);
}
//************************************************************************
// AbstractXMLObjectParser implementation
//************************************************************************
public String getParserDescription() {
return "This element represents an operator that re-aligns a small chunk of an alignment.";
}
public Class getReturnType() {
return IstvanOperator.class;
}
public XMLSyntaxRule[] getSyntaxRules() {
return rules;
}
private XMLSyntaxRule[] rules = new XMLSyntaxRule[]{
AttributeRule.newDoubleRule("weight"),
AttributeRule.newDoubleRule("iP", false, "tuning probability, values near zero resamples entire alignment and near 1.0 resamples single columns."),
AttributeRule.newDoubleRule("exponent", false, "tuning parameter, value of 1.0 samples random alignments, large values (e.g. 2.7) sample alignment peaked around 'best' alignment"),
AttributeRule.newDoubleRule("gapPenalty", false, "tuning parameter, must be negative, large values penalize gaps more in the alignment proposal."),
new ElementRule(TKF91Likelihood.class)
};
}