cache for translation options for speed-up. default to true.

only works for sentence input

git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@1429 1f5c12ca-751b-0410-a591-d2e778427230

8 files changed, 142 insertions, 36 deletions

diff --git a/moses/src/Parameter.cpp b/moses/src/Parameter.cpp
index 2d7b8bf9c..ef6e3986a 100755
--- a/moses/src/Parameter.cpp
+++ b/moses/src/Parameter.cpp
@@ -76,6 +76,7 @@ Parameter::Parameter()
 	AddParam("xml-input", "xi", "allows markup of input with desired translations and probabilities. values can be 'pass-through' (default), 'inclusive', 'exclusive', 'ignore'");
  	AddParam("mbr-scale", "scaling factor to convert log linear score into a probability.");
  	AddParam("decoder-type", "MAP/MBR decoder (default=MAP=0)");
+	AddParam("use-persistent-cache", "cache translation options across sentences (default=true)");
 }
 
 Parameter::~Parameter()
diff --git a/moses/src/StaticData.cpp b/moses/src/StaticData.cpp
index fb2c68b5a..74f2f6e17 100755
--- a/moses/src/StaticData.cpp
+++ b/moses/src/StaticData.cpp
@@ -133,7 +133,18 @@ bool StaticData::LoadData(Parameter *parameter)
 
 	// print all factors of output translations
 	SetBooleanParameter( &m_reportAllFactors, "report-all-factors", false );
-	
+
+	// 
+	if (m_inputType == SentenceInput)
+	{
+		SetBooleanParameter( &m_useTransOptCache, "use-persistent-cache", false );
+	}
+	else
+	{
+		m_useTransOptCache = false;
+	}
+		
+
 	//input factors
 	const vector<string> &inputFactorVector = m_parameter->GetParam("input-factors");
 	for(size_t i=0; i<inputFactorVector.size(); i++) 
@@ -285,6 +296,19 @@ StaticData::~StaticData()
 			}
 	}
 
+	// delete trans opt
+	map<Phrase, std::vector<TranslationOption*> >::iterator iterCache;
+	for (iterCache = m_transOptCache.begin() ; iterCache != m_transOptCache.end() ; ++iterCache)
+	{
+		TranslationOptionList &transOptList = iterCache->second;
+
+		TranslationOptionList::iterator iterTransOpt;
+		for (iterTransOpt = transOptList.begin() ; iterTransOpt != transOptList.end() ; ++iterTransOpt)
+		{
+			delete *iterTransOpt;
+		}
+	}
+
 	RemoveAllInColl(m_reorderModels);
 	
 	// small score producers
@@ -823,3 +847,14 @@ void StaticData::SetWeightsForScoreProducer(const ScoreProducer* sp, const std::
   for (size_t i = begin; i < end; i++)
     m_allWeights[i] = *weightIter++;
 }
+
+const TranslationOptionList* StaticData::FindTransOptListInCache(const Phrase &sourcePhrase) const
+{
+	std::map<Phrase, TranslationOptionList>::const_iterator iter
+			= m_transOptCache.find(sourcePhrase);
+	if (iter == m_transOptCache.end())
+		return NULL;
+
+	return &(iter->second);
+}
+
diff --git a/moses/src/StaticData.h b/moses/src/StaticData.h
index 14c805523..4c1c51551 100755
--- a/moses/src/StaticData.h
+++ b/moses/src/StaticData.h
@@ -111,6 +111,10 @@ protected:
 	XmlInputType m_xmlInputType; //! method for handling sentence XML input
 
   DecoderType m_decoderType; //! MAP or MBR decoder
+
+	bool m_useTransOptCache;
+	mutable std::map<Phrase, TranslationOptionList> m_transOptCache;
+
   float m_mbrScale; //! Scaling factor for computing marginal probability of candidate translation
 	//! constructor. only the 1 static variable can be created
 	StaticData();
@@ -134,6 +138,7 @@ protected:
 	bool LoadLexicalReorderingModel();
 	
 public:
+
 	//! destructor
 	~StaticData();
 	//! return static instance for use like global variable
@@ -337,5 +342,13 @@ public:
 	float GetMBRScale() const {return m_mbrScale;}
 		
 	XmlInputType GetXmlInputType() const { return m_xmlInputType; }
-	
+
+	bool GetUseTransOptCache() const { return m_useTransOptCache; }
+
+	void AddTransOptListToCache(const Phrase &sourcePhrase, const TranslationOptionList &transOptList) const
+	{
+		m_transOptCache[sourcePhrase] = transOptList;
+	}
+
+	const TranslationOptionList* FindTransOptListInCache(const Phrase &sourcePhrase) const;
 };
diff --git a/moses/src/TranslationOption.cpp b/moses/src/TranslationOption.cpp
index b9751bb75..374907b74 100755
--- a/moses/src/TranslationOption.cpp
+++ b/moses/src/TranslationOption.cpp
@@ -89,6 +89,18 @@ TranslationOption::TranslationOption(const TranslationOption &copy)
 , m_reordering(copy.m_reordering)
 {}
 
+TranslationOption::TranslationOption(const TranslationOption &copy, const WordsRange &sourceWordsRange)
+: m_targetPhrase(copy.m_targetPhrase)
+//, m_sourcePhrase(new Phrase(*copy.m_sourcePhrase)) // TODO use when confusion network trans opt for confusion net properly implemented 
+, m_sourcePhrase( (copy.m_sourcePhrase == NULL) ? new Phrase(Input) : new Phrase(*copy.m_sourcePhrase))
+, m_sourceWordsRange(sourceWordsRange)
+, m_totalScore(copy.m_totalScore)
+, m_futureScore(copy.m_futureScore)
+, m_partialScore(copy.m_partialScore)
+, m_scoreBreakdown(copy.m_scoreBreakdown)
+, m_reordering(copy.m_reordering)
+{}
+
 void TranslationOption::MergeNewFeatures(const Phrase& phrase, const ScoreComponentCollection& score, const std::vector<FactorType>& featuresToAdd)
 {
 	assert(phrase.GetSize() == m_targetPhrase.GetSize());
diff --git a/moses/src/TranslationOption.h b/moses/src/TranslationOption.h
index cc74f12a0..a26b3d7a1 100755
--- a/moses/src/TranslationOption.h
+++ b/moses/src/TranslationOption.h
@@ -83,6 +83,9 @@ public:
 	/** copy constructor */
 	TranslationOption(const TranslationOption &copy);
 
+	/** copy constructor, but change words range. used by caching */
+	TranslationOption(const TranslationOption &copy, const WordsRange &sourceWordsRange);
+
 	~TranslationOption()
 	{
 		delete m_sourcePhrase;
diff --git a/moses/src/TranslationOptionCollection.cpp b/moses/src/TranslationOptionCollection.cpp
index 83dd0e608..acc31dedf 100644
--- a/moses/src/TranslationOptionCollection.cpp
+++ b/moses/src/TranslationOptionCollection.cpp
@@ -56,7 +56,6 @@ TranslationOptionCollection::TranslationOptionCollection(InputType const& src, s
 /** destructor, clears out data structures */
 TranslationOptionCollection::~TranslationOptionCollection()
 {
-	// delete all trans opt
 	size_t size = m_source.GetSize();
 	for (size_t startPos = 0 ; startPos < size ; ++startPos)
 	{
@@ -65,6 +64,7 @@ TranslationOptionCollection::~TranslationOptionCollection()
 		 RemoveAllInColl(GetTranslationOptionList(startPos, endPos));
 		}
 	}
+
 	RemoveAllInColl(m_unksrcs);
 }
 
@@ -352,31 +352,47 @@ void TranslationOptionCollection::CreateTranslationOptionsForRange(
 																													 , size_t startPos
 																													 , size_t endPos
 																													 , bool adhereTableLimit)
-{
-
-	if ((StaticData::Instance().GetXmlInputType() != XmlPassThrough) && HasXmlOptionsOverlappingRange(startPos,endPos)) 
+{	
+	if ((StaticData::Instance().GetXmlInputType() != XmlExclusive) || !HasXmlOptionsOverlappingRange(startPos,endPos))
 	{
-		
-		CreateXmlOptionsForRange(startPos, endPos);
+	  Phrase *sourcePhrase = NULL; // can't initialise with substring, in case it's confusion network
 
-	} 
-	
-	if ((StaticData::Instance().GetXmlInputType() != XmlExclusive) || !HasXmlOptionsOverlappingRange(startPos,endPos))
-	{		
-		// partial trans opt stored in here
-		PartialTranslOptColl* oldPtoc = new PartialTranslOptColl;
-		size_t totalEarlyPruned = 0;
-		
-		// initial translation step
-		list <const DecodeStep* >::const_iterator iterStep = decodeStepList.begin();
-		const DecodeStep &decodeStep = **iterStep;
+		// consult persistent (cross-sentence) cache for stored translation options
+		bool skipTransOptCreation = false;
+		if (StaticData::Instance().GetUseTransOptCache()) 
+		{
+		  const WordsRange wordsRange(startPos, endPos);
+		  sourcePhrase = new Phrase(m_source.GetSubString(wordsRange));
+		  
+			const TranslationOptionList *transOptList = StaticData::Instance().FindTransOptListInCache(*sourcePhrase);
+			// is phrase in cache?
+			if (transOptList != NULL) {
+				skipTransOptCreation = true;
+		    TranslationOptionList::const_iterator iterTransOpt;
+		    for (iterTransOpt = transOptList->begin() ; iterTransOpt != transOptList->end() ; ++iterTransOpt)
+				{
+					TranslationOption *transOpt = new TranslationOption(**iterTransOpt, wordsRange);
+					Add(transOpt);
+				}
+			}
+		}
 
-		ProcessInitialTranslation(decodeStep, *oldPtoc
-															, startPos, endPos, adhereTableLimit );
+		if (!skipTransOptCreation)
+		{
+			// partial trans opt stored in here
+			PartialTranslOptColl* oldPtoc = new PartialTranslOptColl;
+			size_t totalEarlyPruned = 0;
+			
+			// initial translation step
+			list <const DecodeStep* >::const_iterator iterStep = decodeStepList.begin();
+			const DecodeStep &decodeStep = **iterStep;
 
-		// do rest of decode steps
-		int indexStep = 0;
-		for (++iterStep ; iterStep != decodeStepList.end() ; ++iterStep) 
+			ProcessInitialTranslation(decodeStep, *oldPtoc
+																, startPos, endPos, adhereTableLimit );
+
+			// do rest of decode steps
+			int indexStep = 0;
+			for (++iterStep ; iterStep != decodeStepList.end() ; ++iterStep) 
 			{
 				const DecodeStep &decodeStep = **iterStep;
 				PartialTranslOptColl* newPtoc = new PartialTranslOptColl;
@@ -399,23 +415,47 @@ void TranslationOptionCollection::CreateTranslationOptionsForRange(
 				oldPtoc = newPtoc;
 				indexStep++;
 			} // for (++iterStep 
-			// add to fully formed translation option list
-		PartialTranslOptColl &lastPartialTranslOptColl	= *oldPtoc;
-		const vector<TranslationOption*>& partTransOptList = lastPartialTranslOptColl.GetList();
-		vector<TranslationOption*>::const_iterator iterColl;
-		for (iterColl = partTransOptList.begin() ; iterColl != partTransOptList.end() ; ++iterColl)
+
+				// add to fully formed translation option list
+			PartialTranslOptColl &lastPartialTranslOptColl	= *oldPtoc;
+			const vector<TranslationOption*>& partTransOptList = lastPartialTranslOptColl.GetList();
+			vector<TranslationOption*>::const_iterator iterColl;
+			for (iterColl = partTransOptList.begin() ; iterColl != partTransOptList.end() ; ++iterColl)
 			{
 				TranslationOption *transOpt = *iterColl;
 				transOpt->CalcScore();
 				Add(transOpt);
 			}
 
-		lastPartialTranslOptColl.DetachAll();
-		totalEarlyPruned += oldPtoc->GetPrunedCount();
-		delete oldPtoc;
-		// TRACE_ERR( "Early translation options pruned: " << totalEarlyPruned << endl);
+			// storing translation options in persistent cache (kept across sentences) 
+			if (StaticData::Instance().GetUseTransOptCache()) 
+			{
+				if (partTransOptList.size() > 0)
+				{
+					vector<TranslationOption*> cachedTransOptList = GetTranslationOptionList(startPos, endPos);
+					vector<TranslationOption*>::iterator iterList;
+					for (size_t i = 0 ; i < cachedTransOptList.size() ; ++i)
+					{
+						cachedTransOptList[i] = new TranslationOption(*cachedTransOptList[i]);
+					}
+	
+					StaticData::Instance().AddTransOptListToCache(*sourcePhrase, cachedTransOptList);
+				}
+				
+				delete sourcePhrase;
+			}
+
+			lastPartialTranslOptColl.DetachAll();
+			totalEarlyPruned += oldPtoc->GetPrunedCount();
+			delete oldPtoc;
+			// TRACE_ERR( "Early translation options pruned: " << totalEarlyPruned << endl);
+		} //if non-exclusive XML or no options for range
+	} // skipTransOptCreation
 
-	} //if non-exclusive XML or no options for range
+	if ((StaticData::Instance().GetXmlInputType() != XmlPassThrough) && HasXmlOptionsOverlappingRange(startPos,endPos)) 
+	{
+		CreateXmlOptionsForRange(startPos, endPos);
+	} 
 
 }
 
diff --git a/moses/src/TranslationOptionCollection.h b/moses/src/TranslationOptionCollection.h
index 012f73cdf..6a0fe7216 100755
--- a/moses/src/TranslationOptionCollection.h
+++ b/moses/src/TranslationOptionCollection.h
@@ -38,8 +38,6 @@ class LMList;
 class FactorMask;
 class Word;
 
-typedef std::vector<TranslationOption*> TranslationOptionList;
-
 /** Contains all phrase translations applicable to current input type (a sentence or confusion network).
  * A key insight into efficient decoding is that various input
  * conditions (trelliss, factored input, normal text, xml markup)
diff --git a/moses/src/TypeDef.h b/moses/src/TypeDef.h
index 2a7714360..0a7738032 100755
--- a/moses/src/TypeDef.h
+++ b/moses/src/TypeDef.h
@@ -23,6 +23,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 
 #include <list>
 #include <limits>
+#include <vector>
 
 #define PROJECT_NAME		"moses"
 
@@ -161,3 +162,6 @@ typedef size_t FactorType;
 typedef uint32_t UINT32;
 #endif
 
+class TranslationOption;
+typedef std::vector<TranslationOption*> TranslationOptionList;
+