commit e8b46bb163003946242a999a93032f431be4dbb0
Author: Tanmai Khanna <khanna.tanmai@gmail.com>
Date:   Fri Jul 19 16:16:00 2019 +0530

    Code to check and use parameters from external XML file | Removed all hardcoded parameters

diff --git a/src/anaphora.cc b/src/anaphora.cc
index dd5c3b4..a721d33 100644
--- a/src/anaphora.cc
+++ b/src/anaphora.cc
@@ -134,7 +134,7 @@ int main(int argc, char **argv)
 					{
 						int retval;
 
-						retval = score_module.add_word(gen_id, sl_form, sl_tags, tl_form); //Give word to Scoring Module
+						retval = score_module.add_word(gen_id, sl_form, sl_tags, tl_form, ref_file.get_parameters()); //Give word to Scoring Module
 						//If retval is 0, nothing will be added in side ref
 
 						//If retval is 1, we call get_antecedent() and add it to ref
diff --git a/src/parse_ref.cc b/src/parse_ref.cc
index 2ac43c0..c3b459f 100644
--- a/src/parse_ref.cc
+++ b/src/parse_ref.cc
@@ -16,8 +16,8 @@ void print_tags(vector< wstring > input)
 {
 	for (int i = 0; i < input.size(); i++) 
 	{
-		wcout << input[i];
-		cout << " ";
+		wcerr << input[i];
+		cerr << " ";
 	}
 }
 
@@ -50,9 +50,9 @@ vector<wstring> ParseRef::parseTags (wstring tags)
 	if(!temptag.empty()) //if any tag remaining
 		temp_tags_list.push_back(temptag);
 
-	print_tags(temp_tags_list);
+	//print_tags(temp_tags_list);
 
-	cout << "\n";
+	//cerr << "\n";
 
 	return temp_tags_list;
 }
@@ -70,7 +70,7 @@ void ParseRef::parseParameterItem (xmlDocPtr doc, xmlNodePtr cur, wstring parame
 	    {
 	    	Attr = xmlGetProp(cur, (const xmlChar *)"tags");
 
-	    	printf("ParameterItem: ");
+	    	//fprintf(stderr, "ParameterItem: ");
 
 	    	temp_tags_list = parseTags(XMLParseUtil::towstring(Attr));
 	    	parameters[parameter_name].push_back(temp_tags_list);
@@ -99,9 +99,9 @@ void ParseRef::parseParameters (xmlDocPtr doc, xmlNodePtr cur)
 		{
 			parameter_name = XMLParseUtil::towstring(cur->name);
 
-			cout << "\n";
-	    	wcout << parameter_name;
-	    	cout << "\n";
+			//cerr << "\n";
+	    	//wcerr << parameter_name;
+	    	//cerr << "\n";
 	    	
 	    	parseParameterItem(doc,cur,parameter_name);
 	    }
@@ -123,7 +123,7 @@ void ParseRef::parseCatItem (xmlDocPtr doc, xmlNodePtr cur, wstring cat_name)
 	    if ((!xmlStrcmp(cur->name, (const xmlChar *)"cat-item"))) 
 	    {
 	    	Attr = xmlGetProp(cur, (const xmlChar *)"tags");
-	    	printf("catItem: ");
+	    	//fprintf(stderr, "catItem: ");
 
 	    	temp_tags_list = parseTags(XMLParseUtil::towstring(Attr));
 		    cats[cat_name].push_back(temp_tags_list);
@@ -150,7 +150,7 @@ void ParseRef::parseCats (xmlDocPtr doc, xmlNodePtr cur)
 	    if ((!xmlStrcmp(cur->name, (const xmlChar *)"def-cat")))
 	    {
 	    	Attr = xmlGetProp(cur, (const xmlChar *)"n");
-	    	printf("catName: %s\n", Attr);  
+	    	//fprintf(stderr, "catName: %s\n", Attr);  
 	    	
 	    	parseCatItem(doc,cur, XMLParseUtil::towstring(Attr));
 	    	xmlFree(Attr);	
@@ -177,8 +177,8 @@ vector<markable_pattern> ParseRef::parsePatternItem (xmlDocPtr doc, xmlNodePtr c
 	    	Attr = xmlGetProp(cur, (const xmlChar *)"n");
 	    	temp.name = XMLParseUtil::towstring(Attr);
 
-	    	wcout << temp.name;
-	    	cout << " ";
+	    	//wcerr << temp.name;
+	    	//cerr << " ";
 
 		    xmlFree(Attr);
 
@@ -187,7 +187,7 @@ vector<markable_pattern> ParseRef::parsePatternItem (xmlDocPtr doc, xmlNodePtr c
 		    if(Attr != NULL)
 		    {
 		    	temp.head = 1;
-		    	printf("[HEAD!]");
+		    	//fprintf(stderr, "[HEAD!]");
 		    }
 		    else
 		    	temp.head = 0;
@@ -209,8 +209,8 @@ void ParseRef::parsePatterns (xmlDocPtr doc, xmlNodePtr cur, wstring markable_na
 
 	vector<markable_pattern> temp_pattern_list;
 
-	wcout << markable_name;
-	cout << "\n";
+	//wcerr << markable_name;
+	//cerr << "\n";
 
 	while (cur != NULL) 
 	{
@@ -222,7 +222,7 @@ void ParseRef::parsePatterns (xmlDocPtr doc, xmlNodePtr cur, wstring markable_na
 
 		cur = cur->next;
 
-		cout << "\n";
+		//cerr << "\n";
 	}
     return;
 }
@@ -237,7 +237,7 @@ void ParseRef::parseMarkables (xmlDocPtr doc, xmlNodePtr cur)
 	    if ((!xmlStrcmp(cur->name, (const xmlChar *)"markable")))
 	    {
 	    	Attr = xmlGetProp(cur, (const xmlChar *)"n");
-	    	printf("MarkableName: ");	
+	    	//fprintf(stderr, "MarkableName: ");	
 	    	
 	    	parsePatterns(doc,cur, XMLParseUtil::towstring(Attr));
 
@@ -303,6 +303,21 @@ void ParseRef::parseDoc(char *docname)
 	return;
 }
 
+unordered_map<wstring, acceptable_tags> ParseRef::get_parameters()
+{
+	return parameters;
+}
+	
+unordered_map<wstring, acceptable_tags> ParseRef::get_cats()
+{
+	return cats;
+}
+
+unordered_map<wstring, acceptable_patterns> ParseRef::get_markables()
+{
+	return markables;
+}
+
 /* //Code for Testing
 int main(int argc, char **argv) 
 {
@@ -310,7 +325,7 @@ int main(int argc, char **argv)
 		
 	if (argc <= 1) 
 	{
-		printf("Usage: %s docname\n", argv[0]);
+		fprintf(stderr, "Usage: %s docname\n", argv[0]);
 		return(0);
 	}
 
diff --git a/src/parse_ref.h b/src/parse_ref.h
index 7a71b67..a1f8385 100644
--- a/src/parse_ref.h
+++ b/src/parse_ref.h
@@ -22,6 +22,8 @@ struct markable_pattern
 
 typedef vector< vector<markable_pattern> > acceptable_patterns;
 
+void print_tags(vector< wstring > input);
+
 class ParseRef
 {
 private:
@@ -43,6 +45,11 @@ public:
 	vector<markable_pattern> parsePatternItem (xmlDocPtr doc, xmlNodePtr cur); 
 
 	vector<wstring> parseTags (wstring tags);
+
+	unordered_map<wstring, acceptable_tags> get_parameters();
+	unordered_map<wstring, acceptable_tags> get_cats();
+
+	unordered_map<wstring, acceptable_patterns> get_markables();
 };
 
 #endif
\ No newline at end of file
diff --git a/src/score.cc b/src/score.cc
index 7b3962f..e4b5de8 100644
--- a/src/score.cc
+++ b/src/score.cc
@@ -1,4 +1,5 @@
 #include "score.h"
+#include "parse_ref.h"
 
 #include<string>
 #include<vector>
@@ -14,15 +15,15 @@ void showq(deque < vector<unique_LU> > gq)
     {
     	vector<unique_LU> temp_sentence = *j;
     	
-    	cout << "\n";
+    	cerr << "\n";
     	for (std::vector<unique_LU>::iterator i = temp_sentence.begin(); i != temp_sentence.end(); ++i)
     	{
-    		wcout << (*i).wordform;
+    		wcerr << (*i).wordform;
     	}
 
-    	cout << "\n";
+    	cerr << "\n";
     } 
-    cout << '\n'; 
+    cerr << '\n'; 
 } 
 
 void clearq(queue < vector<unique_LU> > q)
@@ -52,7 +53,44 @@ int contains_any(vector<wstring> tags, vector<wstring> candidates)
 	return 0; //if no matches
 }
 
-int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wstring > input_pos_tags, wstring input_tl_wordform)
+int check_acceptable_tags(vector<wstring> input_tags, acceptable_tags check_tags)
+{
+	for (acceptable_tags::iterator i = check_tags.begin(); i != check_tags.end(); ++i)
+	{
+		
+		int flag_contains_all = 1;
+
+		vector<wstring> temp_tags = *i;
+
+		for(std::vector<wstring>::iterator j = temp_tags.begin(); j != temp_tags.end(); ++j)
+		{
+
+			if(*j == L"*") //ignore * in the tags list
+				continue;
+
+			if(!contains(input_tags, *j)) //if the required tag is NOT in the input LU tags
+			{
+				flag_contains_all = 0;
+				break;
+			}
+			/*
+			else
+			{
+				cerr << "FoundTag:";
+				wcerr << *j;
+				cerr <<"\n";
+			}
+			*/
+		}
+
+		if(flag_contains_all == 1) //if any tag list fully matched
+			return 1; //else continue to next tag list
+	}
+
+	return 0; //if it didn't return 1 then no tag list was fully matched
+}
+
+int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wstring > input_pos_tags, wstring input_tl_wordform, unordered_map<wstring, acceptable_tags> ref_parameters)
 {
 	unique_LU input_LU = {input_id, input_wordform, input_tl_wordform, input_pos_tags}; //initialise in context with score 0
 
@@ -63,7 +101,7 @@ int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wst
 
 		context.push_back(sentence);
 
-		if(contains(input_LU.pos_tags, L"sent")) //if sentence end (somehow the first LU is a sentence end)
+		if(check_acceptable_tags(input_LU.pos_tags, ref_parameters[L"delimiter"]) ) //if sentence end (somehow the first LU is a sentence end)
 		{
 			vector<unique_LU> new_sentence;
 
@@ -74,7 +112,7 @@ int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wst
 	{
 		context.back().push_back(input_LU); //add word to the latest added sentence in the queue
 
-		if(contains(input_LU.pos_tags, L"sent"))
+		if(check_acceptable_tags(input_LU.pos_tags, ref_parameters[L"delimiter"]) )
 		{
 			vector<unique_LU> new_sentence;
 
@@ -83,9 +121,9 @@ int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wst
 			if(context.size() > 4)
 				context.pop_front(); //remove the earliest added sentence (We only want current and three previous sentences in context)
 		}
-		else if( contains(input_LU.pos_tags, L"det") && contains(input_LU.pos_tags, L"pos")	)
+		else if( check_acceptable_tags(input_LU.pos_tags, ref_parameters[L"anaphor"]) ) //check if tags of current word match with anaphor tags in ref file
 		{
-			apply_indicators(input_LU);
+			apply_indicators(input_LU, ref_parameters);
 			return 1; //To show that something will be added in side ref
 		}
 	}
@@ -93,7 +131,7 @@ int Scoring::add_word(unsigned int input_id, wstring input_wordform, vector< wst
 	return 0; //To show that nothing will be added in side ref
 }
 
-void Scoring::apply_indicators(unique_LU anaphor)
+void Scoring::apply_indicators(unique_LU anaphor, unordered_map<wstring, acceptable_tags> ref_parameters)
 {
 	int distance_marker = 2; //starts from 2 for current sentence and reduces till -1 as we go to previous sentences
 	int temp_score;
@@ -108,7 +146,7 @@ void Scoring::apply_indicators(unique_LU anaphor)
 
 		for (vector<unique_LU>::iterator j = (*i).begin(); j!=(*i).end(); ++j) //read through sentence
 		{
-			if(contains((*j).pos_tags, L"n"))
+			if(check_acceptable_tags((*j).pos_tags, ref_parameters[L"antecedent"]) ) // if it is antecedent (based on external xml file)
 			{
 				temp_score = 0;
 
@@ -135,9 +173,9 @@ void Scoring::apply_indicators(unique_LU anaphor)
 				}
 				else
 				{
-					cout << "\nAgreement Failed for:";
-					wcout << antecedent_LU.wordform;
-					cout << "\n";
+					cerr << "\nAgreement Failed for:";
+					wcerr << antecedent_LU.wordform;
+					cerr << "\n";
 				}
 			}
 		}
@@ -166,9 +204,9 @@ wstring Scoring::get_antecedent()
 
 	for(vector<antecedent>::reverse_iterator it=antecedent_list.rbegin();it!=antecedent_list.rend();++it) //read it in reverse so that we read from furthest to nearest
 	{
-		//cout << "\n" << (*it).LU.id << ": ";
-		//wcout << (*it).LU.wordform;
-		//cout << " : " << (*it).score << "\n";
+		//cerr << "\n" << (*it).LU.id << ": ";
+		//wcerr << (*it).LU.wordform;
+		//cerr << " : " << (*it).score << "\n";
 
 		if((*it).score >= final_antecedent.score) //picking the highest scored and latest added (most recent) antecedent
 			final_antecedent = (*it);
diff --git a/src/score.h b/src/score.h
index 4186363..5336af5 100644
--- a/src/score.h
+++ b/src/score.h
@@ -1,6 +1,8 @@
 #ifndef _MITKOVSCORE_
 #define _MITKOVSCORE_
 
+#include "parse_ref.h"
+
 #include<string>
 #include<vector>
 #include<queue>
@@ -27,6 +29,8 @@ void showq(deque < vector<unique_LU> > gq);
 int contains(vector<wstring> tags, wstring tag);
 int contains_any(vector<wstring> tags, vector<wstring> candidates);
 
+int check_acceptable_tags(vector<wstring> input_tags, acceptable_tags check_tags);
+
 class Scoring
 {
 private:
@@ -34,8 +38,8 @@ private:
 	vector<antecedent> antecedent_list; //A list of antecedents
 
 public:
-	int add_word(unsigned int input_id, wstring input_wordform, vector< wstring > pos_tags, wstring input_tl_wordform);
-	void apply_indicators(unique_LU anaphor);
+	int add_word(unsigned int input_id, wstring input_wordform, vector< wstring > pos_tags, wstring input_tl_wordform, unordered_map<wstring, acceptable_tags> ref_parameters);
+	void apply_indicators(unique_LU anaphor, unordered_map<wstring, acceptable_tags> ref_parameters);
 	int check_agreement(vector<wstring> antecedent_tags, vector<wstring> anaphor_tags);
 	wstring get_antecedent();
 	void clear();