moses-smt/mosesdecoder
1
1
Fork 0
mirror of https://github.com/moses-smt/mosesdecoder.git synced 2024-12-27 22:14:57 +03:00
Code Issues Packages Projects Releases Wiki Activity

create target phrases, not Translation Options, for XML. Can be used in both pb and scfg

Browse Source
This commit is contained in:
Hieu Hoang 2013-07-23 16:46:30 +01:00
parent 361ffa40c4
commit 7380d12af7
2 changed files with 25 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
moses/Sentence.cpp
View File

@ -43,6 +43,11 @@ Sentence::Sentence()
}
}
Sentence::~Sentence()
{
RemoveAllInColl(m_xmlOptions);
}
int Sentence::Read(std::istream& in,const std::vector<FactorType>& factorOrder)
{
const std::string& factorDelimiter = StaticData::Instance().GetFactorDelimiter();
@ -119,12 +124,11 @@ int Sentence::Read(std::istream& in,const std::vector<FactorType>& factorOrder)
// parse XML markup in translation line
//const StaticData &staticData = StaticData::Instance();
std::vector<XmlOption*> xmlOptionsList(0);
std::vector< size_t > xmlWalls;
std::vector< std::pair<size_t, std::string> > placeholders;
if (staticData.GetXmlInputType() != XmlPassThrough) {
if (!ProcessAndStripXMLTags(line, xmlOptionsList, m_reorderingConstraint, xmlWalls, placeholders,
if (!ProcessAndStripXMLTags(line, m_xmlOptions, m_reorderingConstraint, xmlWalls, placeholders,
staticData.GetXmlBrackets().first, staticData.GetXmlBrackets().second)) {
const string msg("Unable to parse XML in line: " + line);
TRACE_ERR(msg << endl);
@ -152,19 +156,15 @@ int Sentence::Read(std::istream& in,const std::vector<FactorType>& factorOrder)
//iterXMLOpts will be empty for XmlIgnore
//look at each column
for(std::vector<XmlOption*>::const_iterator iterXmlOpts = xmlOptionsList.begin();
iterXmlOpts != xmlOptionsList.end(); iterXmlOpts++) {
for(std::vector<XmlOption*>::const_iterator iterXmlOpts = m_xmlOptions.begin();
iterXmlOpts != m_xmlOptions.end(); iterXmlOpts++) {
const XmlOption *xmlOption = *iterXmlOpts;
const WordsRange &range = xmlOption->range;
TranslationOption *transOpt = new TranslationOption(xmlOption->range, xmlOption->targetPhrase);
m_xmlOptionsList.push_back(transOpt);
for(size_t j=transOpt->GetSourceWordsRange().GetStartPos(); j<=transOpt->GetSourceWordsRange().GetEndPos(); j++) {
for(size_t j=range.GetStartPos(); j<=range.GetEndPos(); j++) {
m_xmlCoverageMap[j]=true;
}
delete xmlOption;
}
}
@ -246,10 +246,17 @@ void Sentence::GetXmlTranslationOptions(std::vector <TranslationOption*> &list,
{
//iterate over XmlOptions list, find exact source/target matches
for (std::vector<TranslationOption*>::const_iterator iterXMLOpts = m_xmlOptionsList.begin();
iterXMLOpts != m_xmlOptionsList.end(); iterXMLOpts++) {
if (startPos == (**iterXMLOpts).GetSourceWordsRange().GetStartPos() && endPos == (**iterXMLOpts).GetSourceWordsRange().GetEndPos()) {
list.push_back(*iterXMLOpts);
for (std::vector<XmlOption*>::const_iterator iterXMLOpts = m_xmlOptions.begin();
iterXMLOpts != m_xmlOptions.end(); ++iterXMLOpts) {
const XmlOption &xmlOption = **iterXMLOpts;
const WordsRange &range = xmlOption.range;
if (startPos == range.GetStartPos()
&& endPos == range.GetEndPos()) {
const TargetPhrase &targetPhrase = xmlOption.targetPhrase;
TranslationOption *transOpt = new TranslationOption(range, targetPhrase);
list.push_back(transOpt);
}
}
}

6
moses/Sentence.h
View File

@ -36,6 +36,7 @@ class WordsRange;
class PhraseDictionary;
class TranslationOption;
class TranslationOptionCollection;
class XmlOption;
/**
@ -51,8 +52,8 @@ private:
* Utility method that takes in a string representing an XML tag and the name of the attribute,
* and returns the value of that tag if present, empty string otherwise
*/
std::vector <TranslationOption*> m_xmlOptionsList;
std::vector <bool> m_xmlCoverageMap;
std::vector<XmlOption*> m_xmlOptions;
std::vector <bool> m_xmlCoverageMap;
NonTerminalSet m_defaultLabelSet;
@ -62,6 +63,7 @@ private:
public:
Sentence();
~Sentence();
InputTypeEnum GetType() const {
return SentenceInput;