minor fixes, boolean switches can now be applied without value, e.g. "-t"

git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@803 1f5c12ca-751b-0410-a591-d2e778427230

6 files changed, 63 insertions, 82 deletions

diff --git a/moses/src/Hypothesis.cpp b/moses/src/Hypothesis.cpp
index b08c30e9f..fe0335162 100755
--- a/moses/src/Hypothesis.cpp
+++ b/moses/src/Hypothesis.cpp
@@ -56,7 +56,7 @@ Hypothesis::Hypothesis(InputType const& source, const TargetPhrase &emptyTarget)
 {	// used for initial seeding of trans process	
 	// initialize scores
 	_hash_computed = false;
-	s_HypothesesCreated = 0;
+	s_HypothesesCreated = 1;
 	ResetScore();	
 }
 
diff --git a/moses/src/InputOutput.h b/moses/src/InputOutput.h
index c0759a71b..9ed11d0e5 100755
--- a/moses/src/InputOutput.h
+++ b/moses/src/InputOutput.h
@@ -38,7 +38,7 @@ class InputOutput
  public:
 	InputOutput();
 	virtual InputType* GetInput(InputType*) = 0;
-	virtual void SetOutput(const Hypothesis *hypo, long translationId, bool reportSourceSpan, bool reportAllFactors) = 0;
+	virtual void SetOutput(const Hypothesis *hypo, long translationId, bool reportSegmentation, bool reportAllFactors) = 0;
 	virtual void SetNBest(const LatticePathList &nBestList, long translationId) = 0;
 	virtual ~InputOutput();
 	virtual void Release(InputType*);
diff --git a/moses/src/Parameter.cpp b/moses/src/Parameter.cpp
index a3bfae537..0b54ea9ac 100755
--- a/moses/src/Parameter.cpp
+++ b/moses/src/Parameter.cpp
@@ -53,8 +53,8 @@ Parameter::Parameter()
 	AddParam("n-best-list", "file and size of n-best-list to be generated");
 	AddParam("output-factors", "list of factors in the output");
 	AddParam("phrase-drop-allowed", "da", "if present, allow dropping of source words"); //da = drop any (word); see -du for comparison
-	AddParam("report-all-factors", "?"); //TODO description, please
-	AddParam("report-source-span", "?"); //TODO description, please
+	AddParam("report-all-factors", "report all factors in output, not just first");
+	AddParam("report-segmentation", "t", "report phrase segmentation in the output");
 	AddParam("stack", "s", "maximum stack size for histogram pruning");
 	AddParam("translation-details", "T", "for each best translation hypothesis, print out details about what sourcce spans were used, dropped");
 	AddParam("ttable-file", "location and properties of the translation tables");
@@ -67,50 +67,27 @@ Parameter::Parameter()
 	AddParam("weight-l", "lm", "weight(s) for language models");
 	AddParam("weight-t", "tm", "weights for translation model components");
 	AddParam("weight-w", "w", "weight for word penalty");
-	AddParam("weight-e", "e", "weight for word deletion"); //source word deletion overall weight
-	AddParam("weight-generation", "g", "weight(s) for generation components");
-	AddParam("weight-i", "I", "weight for word insertion");
-	AddParam("mapping", "description of decoding steps");
-	AddParam("n-best-list", "file and size of n-best-list to be generated");
-	AddParam("beam-threshold", "threshold for threshold pruning");
-	AddParam("distortion-limit", "dl", "distortion (reordering) limit in maximum number of words");
-	AddParam("input-factors", "list of factors in the input");
+	AddParam("weight-e", "e", "weight for word deletion"); 
 	AddParam("output-factors", "list if factors in the output");
-	AddParam("mysql", "(deprecated)");
-	AddParam("input-file", "i", "location of the input file to be translated");
 	AddParam("cache-path", "?");
  	AddParam("distortion-file", "source factors (0 if table independent of source), target factors, location of the factorized/lexicalized reordering tables");
  	AddParam("distortion", "configurations for each factorized/lexicalized reordering model.");
-	AddParam("generation-file", "location and properties of the generation table");
-	AddParam("stack", "s", "maximum stack size for histogram pruning");
-	AddParam("verbose", "v", "verbosity level of the logging");
-	AddParam("report-source-span", "?");
-	AddParam("report-all-factors", "?");
-	AddParam("drop-unknown", "du", "drop unknown words instead of copying them");
-	AddParam("inputtype", "text (0) or confusion network (1)");
-	AddParam("translation-details", "T", "for each best translation hypothesis, print out details about what sourcce spans were used, dropped");
-	AddParam("max-trans-opt-per-coverage", "maximum number of translation options per input span (after applying mapping steps)");
-	AddParam("max-partial-trans-opt", "maximum number of partial translation options per input span (during mapping steps)");
-	AddParam("use-distortion-future-costs", "consider expected distortion cost in future cost estimation");
-	AddParam("labeled-n-best-list", "labeled-n-best-list", "print out labels for each weight type in n-best list. default is true");
 }
 
 /** initialize a parameter, sub of constructor */
-PARAM_VEC &Parameter::AddParam(const string &paramName, const string &description)
+void Parameter::AddParam(const string &paramName, const string &description)
 {
 	m_valid[paramName] = true;
 	m_description[paramName] = description;
-	return m_setting[paramName];
 }
 
 /** initialize a parameter (including abbreviation), sub of constructor */
-PARAM_VEC &Parameter::AddParam(const string &paramName, const string &abbrevName, const string &description)
+void Parameter::AddParam(const string &paramName, const string &abbrevName, const string &description)
 {
 	m_valid[paramName] = true;
 	m_valid[abbrevName] = true;
 	m_abbreviation[paramName] = abbrevName;
 	m_description[paramName] = description;
-	return m_setting[paramName];
 }
 
 /** print descriptions of all parameters */
@@ -177,17 +154,16 @@ bool Parameter::LoadParam(int argc, char* argv[])
 
 	// logging of parameters that were set in either config or switch
 	int verbose = 1;
-	if (m_setting["verbose"].size() > 0)
+	if (m_setting.find("verbose") != m_setting.end() &&
+	    m_setting["verbose"].size() > 0)
 	  verbose = Scan<int>(m_setting["verbose"][0]);
 	if (verbose >= 1) { // only if verbose
 	  cerr << "Defined parameters (per moses.ini or switch):" << endl;
 	  for(PARAM_MAP::const_iterator iterParam = m_setting.begin() ; iterParam != m_setting.end(); iterParam++) {
-		if (iterParam->second.size() > 0) {
-			cerr << "\t" << iterParam->first << ": ";
-			for ( size_t i = 0; i < iterParam->second.size(); i++ )
-				cerr << iterParam->second[i] << " ";
-			cerr << endl;
-		}
+	    cerr << "\t" << iterParam->first << ": ";
+	    for ( size_t i = 0; i < iterParam->second.size(); i++ )
+	      cerr << iterParam->second[i] << " ";
+	    cerr << endl;
 	  }
 	}
 
@@ -350,6 +326,7 @@ void Parameter::OverwriteParam(const string &paramSwitch, const string &paramNam
 		return;
 
 	int index = 0;
+	m_setting[paramName]; // defines the parameter, important for boolean switches
 	while (startPos < argc && (!isOption(argv[startPos])))
 	{
 		if (m_setting[paramName].size() > (size_t)index)
diff --git a/moses/src/Parameter.h b/moses/src/Parameter.h
index 3ea5c4be2..321fec7a9 100755
--- a/moses/src/Parameter.h
+++ b/moses/src/Parameter.h
@@ -49,8 +49,8 @@ protected:
 	bool isOption(const char* token);
 	bool Validate();
 
-	PARAM_VEC &AddParam(const std::string &paramName, const std::string &description);
-	PARAM_VEC &AddParam(const std::string &paramName, const std::string &abbrevName, const std::string &description);
+	void AddParam(const std::string &paramName, const std::string &description);
+	void AddParam(const std::string &paramName, const std::string &abbrevName, const std::string &description);
 	
 public:
 	Parameter();
@@ -62,6 +62,11 @@ public:
 	{
 		return m_setting[paramName];
 	}
+	/** check if parameter is defined (either in moses.ini or as switch) */
+	bool isParamSpecified(const std::string &paramName)
+	{
+	  return  m_setting.find( paramName ) != m_setting.end();
+	}
 
 };
 
diff --git a/moses/src/StaticData.cpp b/moses/src/StaticData.cpp
index a2b5313c9..5bd113df0 100755
--- a/moses/src/StaticData.cpp
+++ b/moses/src/StaticData.cpp
@@ -101,27 +101,14 @@ bool StaticData::LoadParameters(int argc, char* argv[])
 		m_nBestSize = 0;
 	}
 	
-	if (m_parameter.GetParam("labeled-n-best-list").size() == 1)
-	{
-		m_labeledNBestList = Scan<bool>( m_parameter.GetParam("labeled-n-best-list")[0]);
-	}
-	else
-	{
-		m_labeledNBestList = true;
-	}
-
+	// include feature names in the n-best list
+	SetBooleanParameter( &m_labeledNBestList, "labeled-n-best-list", true );
 
 	// printing source phrase spans
-	if (m_parameter.GetParam("report-source-span").size() > 0)
-		m_reportSourceSpan = Scan<bool>(m_parameter.GetParam("report-source-span")[0]);
-	else
-        m_reportSourceSpan = false;
+	SetBooleanParameter( &m_reportSegmentation, "report-segmentation", false );
 
 	// print all factors of output translations
-	if (m_parameter.GetParam("report-all-factors").size() > 0)
-		m_reportAllFactors = Scan<bool>(m_parameter.GetParam("report-all-factors")[0]);
-	else
-        m_reportAllFactors = false;
+	SetBooleanParameter( &m_reportAllFactors, "report-all-factors", false );
 
 	//distortion weights
 	 const vector<string> distortionWeights = m_parameter.GetParam("weight-d");	
@@ -158,25 +145,19 @@ bool StaticData::LoadParameters(int argc, char* argv[])
 	}
 
 	//source word deletion
-	if(m_parameter.GetParam("phrase-drop-allowed").size() > 0)
-	{
-		m_wordDeletionEnabled = Scan<bool>(m_parameter.GetParam("phrase-drop-allowed")[0]);
-	}
-	else
-	{
-		m_wordDeletionEnabled = false;
-	}
-	if(m_parameter.GetParam("translation-details").size() > 0) {
-	  m_isDetailedTranslationReportingEnabled = Scan<bool>( m_parameter.GetParam("translation-details")[0]);
-	}
+	SetBooleanParameter( &m_wordDeletionEnabled, "phrase-drop-allowed", false );
 
-	if(m_parameter.GetParam("lmstats").size() > 0) {
-	  m_computeLMBackoffStats = Scan<bool>( m_parameter.GetParam("lmstats")[0]);
-		if (!m_isDetailedTranslationReportingEnabled) {
-			std::cerr << "-lmstats implies -translation-details, enabling" << std::endl;
-			m_isDetailedTranslationReportingEnabled = true;
-		}
+	// additional output
+	SetBooleanParameter( &m_isDetailedTranslationReportingEnabled, 
+			     "translation-details", false );
+
+	SetBooleanParameter( &m_computeLMBackoffStats, "lmstats", false );
+	if (m_computeLMBackoffStats && 
+	    ! m_isDetailedTranslationReportingEnabled) {
+	  std::cerr << "-lmstats implies -translation-details, enabling" << std::endl;
+	  m_isDetailedTranslationReportingEnabled = true;
 	}
+
 	// load Lexical Reordering model
 	const vector<string> &lrFileVector = 
 		m_parameter.GetParam("distortion-file");	
@@ -451,16 +432,33 @@ bool StaticData::LoadParameters(int argc, char* argv[])
 
 	// Unknown Word Processing -- wade
 	//TODO replace this w/general word dropping -- EVH
-	if (m_parameter.GetParam("drop-unknown").size() == 1)
-	  { m_dropUnknown = Scan<bool>( m_parameter.GetParam("drop-unknown")[0]); }
-	else
-	  { m_dropUnknown = 0; }
-
-	//TRACE_ERR("m_dropUnknown: " << m_dropUnknown << endl);
+	SetBooleanParameter( &m_dropUnknown, "drop-unknown", false );
 
 	return true;
 }
 
+void StaticData::SetBooleanParameter( bool *parameter, string parameterName, bool defaultValue ) {
+
+  // default value if nothing is specified
+  *parameter = defaultValue;
+  if (! m_parameter.isParamSpecified( parameterName ) )
+  {
+    return;
+  }
+
+  // if parameter is just specified as, e.g. "-parameter" set it true
+  if (m_parameter.GetParam( parameterName ).size() == 0) 
+  {
+    *parameter = true;
+  }
+
+  // if paramter is specified "-parameter true" or "-parameter false"
+  else if (m_parameter.GetParam( parameterName ).size() == 1) 
+  {
+    *parameter = Scan<bool>( m_parameter.GetParam( parameterName )[0]);
+  }
+}
+
 StaticData::~StaticData()
 {
 	delete m_inputOutput;
diff --git a/moses/src/StaticData.h b/moses/src/StaticData.h
index ffea10979..1dcd526c2 100755
--- a/moses/src/StaticData.h
+++ b/moses/src/StaticData.h
@@ -94,7 +94,7 @@ protected:
 	size_t m_verboseLevel;
 	DistortionScoreProducer *m_distortionScoreProducer;
 	WordPenaltyProducer *m_wpProducer;
-	bool m_reportSourceSpan;
+	bool m_reportSegmentation;
 	bool m_reportAllFactors;
 	bool m_useDistortionFutureCosts;
 	bool m_isDetailedTranslationReportingEnabled;
@@ -113,6 +113,7 @@ public:
 	 * also initialize the Parameter object
 	 */
 	bool LoadParameters(int argc, char* argv[]);
+	void SetBooleanParameter(bool *paramter, string parameterName, bool defaultValue);
 
 	/***
 	 * load not only the main phrase table but also any auxiliary tables that depend on which features are being used
@@ -243,9 +244,9 @@ public:
 	{
 		return m_verboseLevel;
 	}
-	bool GetReportSourceSpan() const
+	bool GetReportSegmentation() const
 	{
-		return m_reportSourceSpan;
+		return m_reportSegmentation;
 	}
 	bool GetReportAllFactors() const
 	{