internal FastFactoredParser(ExhaustivePCFGParser pparser, MLEDependencyGrammar dg, Options op, int numToFind, IGrammarProjection projection, IIndex <string> wordIndex, IIndex <string> tagIndex)
{
this.pparser = pparser;
this.projection = projection;
this.dg = dg;
this.op = op;
this.numToFind = numToFind;
this.wordIndex = wordIndex;
this.tagIndex = tagIndex;
}
protected internal override void TallyRoot(Tree lt, double weight)
{
// this list is in full (not reduced) tag space
IList <IntDependency> deps = MLEDependencyGrammar.TreeToDependencyList(lt, wordIndex, tagIndex);
foreach (IntDependency dependency in deps)
{
dependencyCounter.IncrementCount(dependency, weight);
}
}
public override IDependencyGrammar FormResult()
{
wordIndex.AddToIndex(LexiconConstants.UnknownWord);
MLEDependencyGrammar dg = new MLEDependencyGrammar(tlpParams, directional, useDistance, useCoarseDistance, basicCategoryTagsInDependencyGrammar, op, wordIndex, tagIndex);
foreach (IntDependency dependency in dependencyCounter.KeySet())
{
dg.AddRule(dependency, dependencyCounter.GetCount(dependency));
}
return(dg);
}
/// <summary>Use the DependencyGrammar to score the tree.</summary>
/// <param name="tr">A binarized tree (as returned by the PCFG parser</param>
/// <returns>The score for the tree according to the grammar</returns>
private double DepScoreTree(Tree tr)
{
// log.info("Here's our tree:");
// tr.pennPrint();
// log.info(Trees.toDebugStructureString(tr));
Tree cwtTree = tr.DeepCopy(new LabeledScoredTreeFactory(), new CategoryWordTagFactory());
cwtTree.PercolateHeads(binHeadFinder);
// log.info("Here's what it went to:");
// cwtTree.pennPrint();
IList <IntDependency> deps = MLEDependencyGrammar.TreeToDependencyList(cwtTree, wordIndex, tagIndex);
// log.info("Here's the deps:\n" + deps);
return(dg.ScoreAll(deps));
}
internal FastFactoredParser(ExhaustivePCFGParser pparser, MLEDependencyGrammar dg, Options op, int numToFind, IIndex <string> wordIndex, IIndex <string> tagIndex)
: this(pparser, dg, op, numToFind, new NullGrammarProjection(null, null), wordIndex, tagIndex)
{
}
internal LexicalizedParserQuery(LexicalizedParser parser)
{
this.op = parser.GetOp();
BinaryGrammar bg = parser.bg;
UnaryGrammar ug = parser.ug;
ILexicon lex = parser.lex;
IDependencyGrammar dg = parser.dg;
IIndex <string> stateIndex = parser.stateIndex;
IIndex <string> wordIndex = new DeltaIndex <string>(parser.wordIndex);
IIndex <string> tagIndex = parser.tagIndex;
this.debinarizer = new Debinarizer(op.forceCNF);
this.boundaryRemover = new BoundaryRemover();
if (op.doPCFG)
{
if (op.testOptions.iterativeCKY)
{
pparser = new IterativeCKYPCFGParser(bg, ug, lex, op, stateIndex, wordIndex, tagIndex);
}
else
{
pparser = new ExhaustivePCFGParser(bg, ug, lex, op, stateIndex, wordIndex, tagIndex);
}
}
else
{
pparser = null;
}
if (op.doDep)
{
dg.SetLexicon(lex);
if (!op.testOptions.useFastFactored)
{
dparser = new ExhaustiveDependencyParser(dg, lex, op, wordIndex, tagIndex);
}
else
{
dparser = null;
}
}
else
{
dparser = null;
}
if (op.doDep && op.doPCFG)
{
if (op.testOptions.useFastFactored)
{
MLEDependencyGrammar mledg = (MLEDependencyGrammar)dg;
int numToFind = 1;
if (op.testOptions.printFactoredKGood > 0)
{
numToFind = op.testOptions.printFactoredKGood;
}
bparser = new FastFactoredParser(pparser, mledg, op, numToFind, wordIndex, tagIndex);
}
else
{
IScorer scorer = new TwinScorer(pparser, dparser);
//Scorer scorer = parser;
if (op.testOptions.useN5)
{
bparser = new BiLexPCFGParser.N5BiLexPCFGParser(scorer, pparser, dparser, bg, ug, dg, lex, op, stateIndex, wordIndex, tagIndex);
}
else
{
bparser = new BiLexPCFGParser(scorer, pparser, dparser, bg, ug, dg, lex, op, stateIndex, wordIndex, tagIndex);
}
}
}
else
{
bparser = null;
}
subcategoryStripper = op.tlpParams.SubcategoryStripper();
}