X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=decalc.cpp;h=3b3740bcea6682e14b21300a48e07fdb9f7079a3;hb=1cf188b912d6da8f2cd03dd71cecef664a699c1a;hp=9c84429fe3cc00df92fff4e2ae358ae8553692a4;hpb=8af930b692ff87d1b7031fdd5c9a0b88cdd61245;p=mothur.git

diff --git a/decalc.cpp b/decalc.cpp
index 9c84429..3b3740b 100644
--- a/decalc.cpp
+++ b/decalc.cpp
@@ -12,7 +12,7 @@
 #include "dist.h"
 #include "eachgapdist.h"
 #include "ignoregaps.h"
-
+#include "eachgapdist.h"
 
 //***************************************************************************************************************
 void DeCalculator::setMask(string ms) { 
@@ -683,11 +683,12 @@ float DeCalculator::getCoef(vector<float> obs, vector<float> qav) {
 }
 //***************************************************************************************************************
 //gets closest matches to each end, since chimeras will most likely have different parents on each end
-vector<Sequence*> DeCalculator::findClosest(Sequence* querySeq, vector<Sequence*> db, int& numWanted, vector<int>& indexes) {
+vector<Sequence*> DeCalculator::findClosest(Sequence* querySeq, vector<Sequence*>& thisTemplate, vector<Sequence*>& thisFilteredTemplate, int numWanted, int minSim) {
 	try {
-		indexes.clear();
+		//indexes.clear();
 		
 		vector<Sequence*> seqsMatches;  
+		
 		vector<SeqDist> distsLeft;
 		vector<SeqDist> distsRight;
 		
@@ -743,14 +744,14 @@ vector<Sequence*> DeCalculator::findClosest(Sequence* querySeq, vector<Sequence*
 		Sequence queryRight(querySeq->getName(), rightQuery);
 //cout << querySeq->getName() << '\t' << leftSpot << '\t' << rightSpot << '\t' << firstBaseSpot << '\t' << lastBaseSpot << endl;
 //cout << queryUnAligned.length() << '\t' << queryLeft.getUnaligned().length() << '\t' << queryRight.getUnaligned().length() << endl;
-		for(int j = 0; j < db.size(); j++){
+		for(int j = 0; j < thisFilteredTemplate.size(); j++){
 			
-			string dbAligned = db[j]->getAligned();
+			string dbAligned = thisFilteredTemplate[j]->getAligned();
 			string leftDB = dbAligned.substr(firstBaseSpot, (leftSpot-firstBaseSpot+1)); //first 1/3 of the sequence
 			string rightDB = dbAligned.substr(rightSpot, (lastBaseSpot-rightSpot)); //last 1/3 of the sequence
 			
-			Sequence dbLeft(db[j]->getName(), leftDB);
-			Sequence dbRight(db[j]->getName(), rightDB);
+			Sequence dbLeft(thisFilteredTemplate[j]->getName(), leftDB);
+			Sequence dbRight(thisFilteredTemplate[j]->getName(), rightDB);
 
 			distcalculator->calcDist(queryLeft, dbLeft);
 			float distLeft = distcalculator->getDist();
@@ -759,14 +760,14 @@ vector<Sequence*> DeCalculator::findClosest(Sequence* querySeq, vector<Sequence*
 			float distRight = distcalculator->getDist();
 			
 			SeqDist subjectLeft;
-			subjectLeft.seq = db[j];
+			subjectLeft.seq = NULL;
 			subjectLeft.dist = distLeft;
 			subjectLeft.index = j;
 			
 			distsLeft.push_back(subjectLeft);
 			
 			SeqDist subjectRight;
-			subjectRight.seq = db[j];
+			subjectRight.seq = NULL;
 			subjectRight.dist = distRight;
 			subjectRight.index = j;
 			
@@ -787,51 +788,62 @@ vector<Sequence*> DeCalculator::findClosest(Sequence* querySeq, vector<Sequence*
 		vector<SeqDist> dists;
 		float lastRight = distsRight[0].dist;
 		float lastLeft = distsLeft[0].dist;
-		int lasti = 0;
-		for (int i = 0; i < distsLeft.size(); i++) {
+		//int lasti = 0;
+		for (int i = 0; i < numWanted+1; i++) {
+			
+			if (m->control_pressed) { return seqsMatches; }
+			
 			//add left if you havent already
-			it = seen.find(distsLeft[i].seq->getName());
+			it = seen.find(thisTemplate[distsLeft[i].index]->getName());
 			if (it == seen.end()) {  
 				dists.push_back(distsLeft[i]);
-				seen[distsLeft[i].seq->getName()] = distsLeft[i].seq->getName();
+				seen[thisTemplate[distsLeft[i].index]->getName()] = thisTemplate[distsLeft[i].index]->getName();
 				lastLeft =  distsLeft[i].dist;
+//				cout << "loop-left\t" << db[distsLeft[i].index]->getName() << '\t' << distsLeft[i].dist << endl;
 			}
 
 			//add right if you havent already
-			it = seen.find(distsRight[i].seq->getName());
+			it = seen.find(thisTemplate[distsRight[i].index]->getName());
 			if (it == seen.end()) {  
 				dists.push_back(distsRight[i]);
-				seen[distsRight[i].seq->getName()] = distsRight[i].seq->getName();
+				seen[thisTemplate[distsRight[i].index]->getName()] = thisTemplate[distsRight[i].index]->getName();
 				lastRight =  distsRight[i].dist;
+//				cout << "loop-right\t" << db[distsRight[i].index]->getName() << '\t' << distsRight[i].dist << endl;
 			}
 			
-			if (dists.size() > numWanted) { lasti = i; break; } //you have enough results
 		}
 		
-		//add in dups
-		lasti++;
-		int i = lasti;
-		while (i < distsLeft.size()) {  
-			if (distsLeft[i].dist == lastLeft) {  dists.push_back(distsLeft[i]);  numWanted++; }
-			else { break; }
-			i++;
+		//are we still above the minimum similarity cutoff
+		if ((lastLeft >= minSim) || (lastRight >= minSim)) {
+			//add in ties from left
+			int i = numWanted;
+			while (i < distsLeft.size()) {  
+				if (distsLeft[i].dist == lastLeft) {  dists.push_back(distsLeft[i]);  }
+				else { break; }
+				i++;
+			}
+			
+			//add in ties from right
+			i = numWanted;
+			while (i < distsRight.size()) {  
+				if (distsRight[i].dist == lastRight) {  dists.push_back(distsRight[i]);  }
+				else { break; }
+				i++;
+			}
 		}
 		
-		i = lasti;
-		while (i < distsRight.size()) {  
-			if (distsRight[i].dist == lastRight) {  dists.push_back(distsRight[i]);  numWanted++; }
-			else { break; }
-			i++;
-		}
 
-		if (numWanted > dists.size()) { m->mothurOut("numwanted is larger than the number of template sequences, adjusting numwanted."); m->mothurOutEndLine(); numWanted = dists.size(); }
 
-//cout << numWanted << endl;
-		for (int i = 0; i < numWanted; i++) {
-//cout << dists[i].seq->getName() << '\t' << dists[i].dist << endl;
-			Sequence* temp = new Sequence(dists[i].seq->getName(), dists[i].seq->getAligned()); //have to make a copy so you can trim and filter without stepping on eachother.
-			seqsMatches.push_back(temp);
-			indexes.push_back(dists[i].index);
+		//cout << numWanted << endl;
+		for (int i = 0; i < dists.size(); i++) {
+//			cout << db[dists[i].index]->getName() << '\t' << dists[i].dist << endl;
+
+			if ((thisTemplate[dists[i].index]->getName() != querySeq->getName()) && (dists[i].dist >= minSim)) {
+				Sequence* temp = new Sequence(thisTemplate[dists[i].index]->getName(), thisTemplate[dists[i].index]->getAligned()); //have to make a copy so you can trim and filter without stepping on eachother.
+			
+				seqsMatches.push_back(temp);
+			}
+
 		}
 		
 		return seqsMatches;
@@ -951,7 +963,6 @@ map<int, int> DeCalculator::trimSeqs(Sequence* query, vector<Sequence*> topMatch
 		map<int, int> trimmedPos;
 		//check to make sure that is not whole seq
 		if ((rearPos - frontPos - 1) <= 0) {  
-			m->mothurOut("[ERROR]: when I trim your sequences, the entire sequence is trimmed."); m->mothurOutEndLine();  
 			query->setAligned("");
 			//trim topMatches
 			for (int i = 0; i < topMatches.size(); i++) {