mirror of
https://github.com/moses-smt/mosesdecoder.git
synced 2024-09-19 23:27:46 +03:00
99 lines
3.3 KiB
C++
99 lines
3.3 KiB
C++
#include "TreeStructureFeature.h"
|
|
#include "moses/StaticData.h"
|
|
#include "moses/ScoreComponentCollection.h"
|
|
#include "moses/ChartHypothesis.h"
|
|
#include <vector>
|
|
#include "moses/PP/TreeStructurePhraseProperty.h"
|
|
|
|
namespace Moses
|
|
{
|
|
|
|
void TreeStructureFeature::Load()
|
|
{
|
|
|
|
// syntactic constraints can be hooked in here.
|
|
m_constraints = NULL;
|
|
m_labelset = NULL;
|
|
|
|
StaticData &staticData = StaticData::InstanceNonConst();
|
|
staticData.SetTreeStructure(this);
|
|
}
|
|
|
|
|
|
// define NT labels (ints) that are mapped from strings for quicker comparison.
|
|
void TreeStructureFeature::AddNTLabels(TreePointer root) const
|
|
{
|
|
std::string label = root->GetLabel();
|
|
|
|
if (root->IsTerminal()) {
|
|
return;
|
|
}
|
|
|
|
std::map<std::string, NTLabel>::const_iterator it = m_labelset->string_to_label.find(label);
|
|
if (it != m_labelset->string_to_label.end()) {
|
|
root->SetNTLabel(it->second);
|
|
}
|
|
|
|
std::vector<TreePointer> children = root->GetChildren();
|
|
for (std::vector<TreePointer>::const_iterator it2 = children.begin(); it2 != children.end(); ++it2) {
|
|
AddNTLabels(*it2);
|
|
}
|
|
}
|
|
|
|
FFState* TreeStructureFeature::EvaluateWhenApplied(const ChartHypothesis& cur_hypo
|
|
, int featureID /* used to index the state in the previous hypotheses */
|
|
, ScoreComponentCollection* accumulator) const
|
|
{
|
|
if (const PhraseProperty *property = cur_hypo.GetCurrTargetPhrase().GetProperty("Tree")) {
|
|
const std::string *tree = property->GetValueString();
|
|
TreePointer mytree (boost::make_shared<InternalTree>(*tree));
|
|
|
|
if (m_labelset) {
|
|
AddNTLabels(mytree);
|
|
}
|
|
|
|
//get subtrees (in target order)
|
|
std::vector<TreePointer> previous_trees;
|
|
for (size_t pos = 0; pos < cur_hypo.GetCurrTargetPhrase().GetSize(); ++pos) {
|
|
const Word &word = cur_hypo.GetCurrTargetPhrase().GetWord(pos);
|
|
if (word.IsNonTerminal()) {
|
|
size_t nonTermInd = cur_hypo.GetCurrTargetPhrase().GetAlignNonTerm().GetNonTermIndexMap()[pos];
|
|
const ChartHypothesis *prevHypo = cur_hypo.GetPrevHypo(nonTermInd);
|
|
const TreeState* prev = dynamic_cast<const TreeState*>(prevHypo->GetFFState(featureID));
|
|
const TreePointer prev_tree = prev->GetTree();
|
|
previous_trees.push_back(prev_tree);
|
|
}
|
|
}
|
|
|
|
if (m_constraints) {
|
|
m_constraints->SyntacticRules(mytree, previous_trees, this, accumulator);
|
|
}
|
|
mytree->Combine(previous_trees);
|
|
|
|
bool full_sentence = (mytree->GetChildren().back()->GetLabel() == "</s>" || (mytree->GetChildren().back()->GetLabel() == "SEND" && mytree->GetChildren().back()->GetChildren().back()->GetLabel() == "</s>"));
|
|
if (m_binarized && full_sentence) {
|
|
mytree->Unbinarize();
|
|
}
|
|
|
|
return new TreeState(mytree);
|
|
} else {
|
|
UTIL_THROW2("Error: TreeStructureFeature active, but no internal tree structure found");
|
|
}
|
|
|
|
}
|
|
|
|
void TreeStructureFeature::SetParameter(const std::string& key, const std::string& value)
|
|
{
|
|
std::cerr << "setting: " << this->GetScoreProducerDescription() << " - " << key << "\n";
|
|
if (key == "tuneable") {
|
|
m_tuneable = Scan<bool>(value);
|
|
} else if (key == "filterable") { //ignore
|
|
} else if (key == "binarized") { // if trees have been binarized before learning translation model; output unbinarized trees
|
|
m_binarized = true;
|
|
} else {
|
|
UTIL_THROW(util::Exception, "Unknown argument " << key << "=" << value);
|
|
}
|
|
}
|
|
|
|
}
|