+/*
+ * shhher.cpp
+ * Mothur
+ *
+ * Created by Pat Schloss on 12/27/10.
+ * Copyright 2010 Schloss Lab. All rights reserved.
+ *
+ */
+
+#include "shhhercommand.h"
+
+#include "readcolumn.h"
+#include "readmatrix.hpp"
+#include "rabundvector.hpp"
+#include "sabundvector.hpp"
+#include "listvector.hpp"
+#include "cluster.hpp"
+#include "sparsematrix.hpp"
+
+//**********************************************************************************************************************
+
+#define NUMBINS 1000
+#define HOMOPS 10
+#define MIN_COUNT 0.1
+#define MIN_WEIGHT 0.1
+#define MIN_TAU 0.0001
+#define MIN_ITER 10
+
+//**********************************************************************************************************************
+
+vector<string> ShhherCommand::getValidParameters(){
+ try {
+ string Array[] = {
+ "file", "flow", "lookup", "cutoff", "sigma", "outputdir","inputdir", "processors"
+ };
+
+ vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getValidParameters");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+ShhherCommand::ShhherCommand(){
+ try {
+ abort = true;
+
+ //initialize outputTypes
+ vector<string> tempOutNames;
+ outputTypes["pn.dist"] = tempOutNames;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "ShhherCommand");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+vector<string> ShhherCommand::getRequiredParameters(){
+ try {
+ string Array[] = {"flow"};
+ vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getRequiredParameters");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+vector<string> ShhherCommand::getRequiredFiles(){
+ try {
+ vector<string> myArray;
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getRequiredFiles");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+ShhherCommand::ShhherCommand(string option) {
+ try {
+
+#ifdef USE_MPI
+ MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
+ MPI_Comm_size(MPI_COMM_WORLD, &ncpus);
+
+ if(pid == 0){
+#endif
+
+
+ abort = false;
+
+
+ //allow user to run help
+ if(option == "help") { help(); abort = true; }
+
+ else {
+
+ //valid paramters for this command
+ string AlignArray[] = {
+ "file", "flow", "lookup", "cutoff", "sigma", "outputdir","inputdir", "processors"
+ };
+
+ vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
+
+ OptionParser parser(option);
+ map<string,string> parameters = parser.getParameters();
+
+ ValidParameters validParameter;
+ map<string,string>::iterator it;
+
+ //check to make sure all parameters are valid for command
+ for (it = parameters.begin(); it != parameters.end(); it++) {
+ if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
+ }
+
+ //initialize outputTypes
+ vector<string> tempOutNames;
+ outputTypes["pn.dist"] = tempOutNames;
+ // outputTypes["fasta"] = tempOutNames;
+
+ //if the user changes the input directory command factory will send this info to us in the output parameter
+ string inputDir = validParameter.validFile(parameters, "inputdir", false);
+ if (inputDir == "not found"){ inputDir = ""; }
+ else {
+ string path;
+ it = parameters.find("flow");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = m->hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["flow"] = inputDir + it->second; }
+ }
+
+ it = parameters.find("lookup");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = m->hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["lookup"] = inputDir + it->second; }
+ }
+
+ it = parameters.find("file");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = m->hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["file"] = inputDir + it->second; }
+ }
+ }
+
+
+ //check for required parameters
+ flowFileName = validParameter.validFile(parameters, "flow", true);
+ flowFilesFileName = validParameter.validFile(parameters, "file", true);
+ if (flowFileName == "not found" && flowFilesFileName == "not found") {
+ m->mothurOut("values for either flow or file must be provided for the shhh.seqs command.");
+ m->mothurOutEndLine();
+ abort = true;
+ }
+ else if (flowFileName == "not open" || flowFilesFileName == "not open") { abort = true; }
+
+ //if the user changes the output directory command factory will send this info to us in the output parameter
+ outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
+ outputDir = "";
+ outputDir += m->hasPath(flowFileName); //if user entered a file with a path then preserve it
+ }
+
+
+ //check for optional parameter and set defaults
+ // ...at some point should added some additional type checking...
+ string temp;
+ temp = validParameter.validFile(parameters, "lookup", true);
+ if (temp == "not found") { lookupFileName = "LookUp_E123.pat"; }
+ else if(temp == "not open") { abort = true; }
+ else { lookupFileName = temp; }
+
+ temp = validParameter.validFile(parameters, "processors", false);if (temp == "not found"){ temp = "1"; }
+ convert(temp, processors);
+
+ temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found"){ temp = "0.01"; }
+ convert(temp, cutoff);
+
+ temp = validParameter.validFile(parameters, "mindelta", false); if (temp == "not found"){ temp = "0.000001"; }
+ convert(temp, minDelta);
+
+ temp = validParameter.validFile(parameters, "maxiter", false); if (temp == "not found"){ temp = "1000"; }
+ convert(temp, maxIters);
+
+ temp = validParameter.validFile(parameters, "sigma", false);if (temp == "not found") { temp = "60"; }
+ convert(temp, sigma);
+
+ globaldata = GlobalData::getInstance();
+ }
+
+#ifdef USE_MPI
+ }
+#endif
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "ShhherCommand");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+ShhherCommand::~ShhherCommand(){}
+
+//**********************************************************************************************************************
+
+void ShhherCommand::help(){
+ try {
+ m->mothurOut("The shhher command reads a file containing flowgrams and creates a file of corrected sequences.\n");
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "help");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+#ifdef USE_MPI
+int ShhherCommand::execute(){
+ try {
+ int tag = 1976;
+ MPI_Status status;
+
+ double begClock = clock();
+ unsigned long int begTime = time(NULL);
+
+ cout.setf(ios::fixed, ios::floatfield);
+ cout.setf(ios::showpoint);
+ cout << setprecision(2);
+
+
+ if(pid == 0){
+
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&abort, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ }
+ if(abort == 1){ return 0; }
+
+ processors = ncpus;
+
+ cout << "\nGetting preliminary data..." << endl;
+ getSingleLookUp();
+ getJointLookUp();
+
+ vector<string> flowFileVector;
+ if(flowFilesFileName != "not found"){
+ string fName;
+
+ ifstream flowFilesFile;
+ m->openInputFile(flowFilesFileName, flowFilesFile);
+ while(flowFilesFile){
+ flowFilesFile >> fName;
+ flowFileVector.push_back(fName);
+ m->gobble(flowFilesFile);
+ }
+ }
+ else{
+ flowFileVector.push_back(flowFileName);
+ }
+ int numFiles = flowFileVector.size();
+
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&numFiles, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ }
+
+ for(int i=0;i<numFiles;i++){
+ flowFileName = flowFileVector[i];
+
+ cout << "\n>>>>>\tProcessing " << flowFileName << " (file " << i+1 << " of " << numFiles << ")\t<<<<<" << endl;
+ cout << "Reading flowgrams..." << endl;
+ getFlowData();
+ cout << "Identifying unique flowgrams..." << endl;
+ getUniques();
+
+ cout << "Calculating distances between flowgrams..." << endl;
+ char fileName[1024];
+ strcpy(fileName, flowFileName.c_str());
+
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&fileName[0], 1024, MPI_CHAR, i, tag, MPI_COMM_WORLD);
+
+ MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&numUniques, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&numFlowCells, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&flowDataIntI[0], numSeqs * numFlowCells, MPI_SHORT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&flowDataPrI[0], numSeqs * numFlowCells, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&mapUniqueToSeq[0], numSeqs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&mapSeqToUnique[0], numSeqs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&lengths[0], numSeqs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&jointLookUp[0], NUMBINS * NUMBINS, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&cutoff, 1, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ }
+
+ string distFileName = flowDistMPI(0, int(sqrt(1.0/float(ncpus)) * numUniques));
+
+ int done;
+ for(int i=1;i<ncpus;i++){
+ MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
+
+ m->appendFiles((distFileName + ".temp." + toString(i)), distFileName);
+ remove((distFileName + ".temp." + toString(i)).c_str());
+ }
+
+ string namesFileName = createNamesFile();
+
+ cout << "\nClustering flowgrams..." << endl;
+ string listFileName = cluster(distFileName, namesFileName);
+ // string listFileName = "PriestPot_C7.pn.list";
+ // string listFileName = "test.mock_rep3.v69.pn.list";
+
+ getOTUData(listFileName);
+ initPyroCluster();
+
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&numOTUs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&singleLookUp[0], singleLookUp.size(), MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&uniqueFlowgrams[0], numFlowCells * numUniques, MPI_SHORT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&sigma, 1, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ }
+
+
+ double maxDelta = 0;
+ int iter = 0;
+
+ int numOTUsOnCPU = numOTUs / ncpus;
+ int numSeqsOnCPU = numSeqs / ncpus;
+
+ cout << "\nDenoising flowgrams..." << endl;
+ cout << "iter\tmaxDelta\tnLL\t\tcycletime" << endl;
+
+ while((maxIters == 0 && maxDelta > minDelta) || iter < MIN_ITER || (maxDelta > minDelta && iter < maxIters)){
+
+ double cycClock = clock();
+ unsigned long int cycTime = time(NULL);
+ fill();
+
+ int total = singleTau.size();
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&total, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&change[0], numOTUs, MPI_SHORT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(¢roids[0], numOTUs, MPI_INT, i, tag, MPI_COMM_WORLD);
+
+ MPI_Send(&singleTau[0], total, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&seqNumber[0], total, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&seqIndex[0], total, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&nSeqsPerOTU[0], numOTUs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&cumNumSeqs[0], numOTUs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ }
+
+ calcCentroidsDriver(0, numOTUsOnCPU);
+
+ for(int i=1;i<ncpus;i++){
+ int otuStart = i * numOTUs / ncpus;
+ int otuStop = (i + 1) * numOTUs / ncpus;
+
+ vector<int> tempCentroids(numOTUs, 0);
+ vector<short> tempChange(numOTUs, 0);
+
+ MPI_Recv(&tempCentroids[0], numOTUs, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&tempChange[0], numOTUs, MPI_SHORT, i, tag, MPI_COMM_WORLD, &status);
+
+ for(int j=otuStart;j<otuStop;j++){
+ centroids[j] = tempCentroids[j];
+ change[j] = tempChange[j];
+ }
+ }
+
+ maxDelta = getNewWeights();
+ double nLL = getLikelihood();
+ checkCentroids();
+
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(¢roids[0], numOTUs, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&weight[0], numOTUs, MPI_DOUBLE, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&change[0], numOTUs, MPI_SHORT, i, tag, MPI_COMM_WORLD);
+ }
+
+ calcNewDistancesParent(0, numSeqsOnCPU);
+
+ total = singleTau.size();
+
+ for(int i=1;i<ncpus;i++){
+ int childTotal;
+ int seqStart = i * numSeqs / ncpus;
+ int seqStop = (i + 1) * numSeqs / ncpus;
+
+ MPI_Recv(&childTotal, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
+
+ vector<int> childSeqIndex(childTotal, 0);
+ vector<double> childSingleTau(childTotal, 0);
+ vector<double> childDist(numSeqs * numOTUs, 0);
+ vector<int> otuIndex(childTotal, 0);
+
+ MPI_Recv(&childSeqIndex[0], childTotal, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&childSingleTau[0], childTotal, MPI_DOUBLE, i, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&childDist[0], numOTUs * numSeqs, MPI_DOUBLE, i, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&otuIndex[0], childTotal, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
+
+ int oldTotal = total;
+ total += childTotal;
+ singleTau.resize(total, 0);
+ seqIndex.resize(total, 0);
+ seqNumber.resize(total, 0);
+
+ int childIndex = 0;
+
+ for(int j=oldTotal;j<total;j++){
+ int otuI = otuIndex[childIndex];
+ int seqI = childSeqIndex[childIndex];
+
+ singleTau[j] = childSingleTau[childIndex];
+
+ aaP[otuI][nSeqsPerOTU[otuI]] = j;
+ aaI[otuI][nSeqsPerOTU[otuI]] = seqI;
+ nSeqsPerOTU[otuI]++;
+ childIndex++;
+ }
+
+ int index = seqStart * numOTUs;
+ for(int j=seqStart;j<seqStop;j++){
+ for(int k=0;k<numOTUs;k++){
+ dist[index] = childDist[index];
+ index++;
+ }
+ }
+ }
+
+ iter++;
+
+ cout << iter << '\t' << maxDelta << '\t' << setprecision(2) << nLL << '\t' << time(NULL) - cycTime << '\t' << setprecision(6) << (clock() - cycClock)/(double)CLOCKS_PER_SEC << endl;
+
+ if((maxIters == 0 && maxDelta > minDelta) || iter < MIN_ITER || (maxDelta > minDelta && iter < maxIters)){
+ int live = 1;
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&live, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ }
+ }
+ else{
+ int live = 0;
+ for(int i=1;i<ncpus;i++){
+ MPI_Send(&live, 1, MPI_INT, i, tag, MPI_COMM_WORLD); //send kill command
+ }
+ }
+
+ }
+
+ cout << "\nFinalizing..." << endl;
+ fill();
+ setOTUs();
+ vector<int> otuCounts(numOTUs, 0);
+ for(int i=0;i<numSeqs;i++) { otuCounts[otuData[i]]++; }
+ calcCentroidsDriver(0, numOTUs);
+ writeQualities(otuCounts);
+ writeSequences(otuCounts);
+ writeNames(otuCounts);
+ writeClusters(otuCounts);
+ writeGroups();
+
+ remove(distFileName.c_str());
+ remove(namesFileName.c_str());
+ remove(listFileName.c_str());
+
+ cout << "Total time to process " << flowFileName << ":\t" << time(NULL) - begTime << '\t' << setprecision(6) << (clock() - begClock)/(double)CLOCKS_PER_SEC << endl;
+ }
+
+ }
+ else{
+ int abort = 1;
+ bool live = 1;
+
+ MPI_Recv(&abort, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ if(abort){ return 0; }
+
+ int numFiles;
+ MPI_Recv(&numFiles, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
+ for(int i=0;i<numFiles;i++){
+ //Now into the pyrodist part
+ char fileName[1024];
+ MPI_Recv(&fileName, 1024, MPI_CHAR, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&numUniques, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&numFlowCells, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
+ flowDataIntI.resize(numSeqs * numFlowCells);
+ flowDataPrI.resize(numSeqs * numFlowCells);
+ mapUniqueToSeq.resize(numSeqs);
+ mapSeqToUnique.resize(numSeqs);
+ lengths.resize(numSeqs);
+ jointLookUp.resize(NUMBINS * NUMBINS);
+
+ MPI_Recv(&flowDataIntI[0], numSeqs * numFlowCells, MPI_SHORT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&flowDataPrI[0], numSeqs * numFlowCells, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&mapUniqueToSeq[0], numSeqs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&mapSeqToUnique[0], numSeqs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&lengths[0], numSeqs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&jointLookUp[0], NUMBINS * NUMBINS, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&cutoff, 1, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+
+ flowFileName = string(fileName);
+ int flowDistStart = int(sqrt(float(pid)/float(ncpus)) * numUniques);
+ int flowDistEnd = int(sqrt(float(pid+1)/float(ncpus)) * numUniques);
+
+ string distanceStringChild = flowDistMPI(flowDistStart, flowDistEnd);
+
+ int done = 1;
+ MPI_Send(&done, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
+
+ //Now into the pyronoise part
+ MPI_Recv(&numOTUs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
+ singleLookUp.resize(HOMOPS * NUMBINS);
+ uniqueFlowgrams.resize(numUniques * numFlowCells);
+ weight.resize(numOTUs);
+ centroids.resize(numOTUs);
+ change.resize(numOTUs);
+ dist.assign(numOTUs * numSeqs, 0);
+ nSeqsPerOTU.resize(numOTUs);
+ cumNumSeqs.resize(numOTUs);
+
+ MPI_Recv(&singleLookUp[0], singleLookUp.size(), MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&uniqueFlowgrams[0], uniqueFlowgrams.size(), MPI_SHORT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&sigma, 1, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+
+ int startOTU = pid * numOTUs / ncpus;
+ int endOTU = (pid + 1) * numOTUs / ncpus;
+
+ int startSeq = pid * numSeqs / ncpus;
+ int endSeq = (pid + 1) * numSeqs /ncpus;
+
+ int total;
+
+ while(live){
+
+ MPI_Recv(&total, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ singleTau.assign(total, 0.0000);
+ seqNumber.assign(total, 0);
+ seqIndex.assign(total, 0);
+
+ MPI_Recv(&change[0], numOTUs, MPI_SHORT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(¢roids[0], numOTUs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&singleTau[0], total, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&seqNumber[0], total, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&seqIndex[0], total, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&nSeqsPerOTU[0], total, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&cumNumSeqs[0], numOTUs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
+ calcCentroidsDriver(startOTU, endOTU);
+
+ MPI_Send(¢roids[0], numOTUs, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(&change[0], numOTUs, MPI_SHORT, 0, tag, MPI_COMM_WORLD);
+
+
+ MPI_Recv(¢roids[0], numOTUs, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&weight[0], numOTUs, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&change[0], numOTUs, MPI_SHORT, 0, tag, MPI_COMM_WORLD, &status);
+
+ vector<int> otuIndex(total, 0);
+ calcNewDistancesChildMPI(startSeq, endSeq, otuIndex);
+ total = otuIndex.size();
+
+ MPI_Send(&total, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(&seqIndex[0], total, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(&singleTau[0], total, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(&dist[0], numOTUs * numSeqs, MPI_DOUBLE, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(&otuIndex[0], total, MPI_INT, 0, tag, MPI_COMM_WORLD);
+
+ MPI_Recv(&live, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ }
+ }
+ }
+
+ MPI_Barrier(MPI_COMM_WORLD);
+ return 0;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "execute");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+string ShhherCommand::flowDistMPI(int startSeq, int stopSeq){
+ try{
+ ostringstream outStream;
+ outStream.setf(ios::fixed, ios::floatfield);
+ outStream.setf(ios::dec, ios::basefield);
+ outStream.setf(ios::showpoint);
+ outStream.precision(6);
+
+ int begTime = time(NULL);
+ double begClock = clock();
+
+ for(int i=startSeq;i<stopSeq;i++){
+ for(int j=0;j<i;j++){
+ float flowDistance = calcPairwiseDist(mapUniqueToSeq[i], mapUniqueToSeq[j]);
+
+ if(flowDistance < 1e-6){
+ outStream << mapUniqueToSeq[i] << '\t' << mapUniqueToSeq[j] << '\t' << 0.000000 << endl;
+ }
+ else if(flowDistance <= cutoff){
+ outStream << mapUniqueToSeq[i] << '\t' << mapUniqueToSeq[j] << '\t' << flowDistance << endl;
+ }
+ }
+ if(i % 100 == 0){
+ cout << i << "\t" << (time(NULL) - begTime) << "\t" << (clock()-begClock)/CLOCKS_PER_SEC << endl;
+ }
+ }
+ cout << stopSeq << "\t" << (time(NULL) - begTime) << "\t" << (clock()-begClock)/CLOCKS_PER_SEC << endl;
+
+ string fDistFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.dist";
+ if(pid != 0){ fDistFileName += ".temp." + toString(pid); }
+
+ ofstream distFile(fDistFileName.c_str());
+ distFile << outStream.str();
+ distFile.close();
+
+ return fDistFileName;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "flowDistParentFork");
+ exit(1);
+ }
+}
+
+#else
+//**********************************************************************************************************************
+
+int ShhherCommand::execute(){
+ try {
+ if (abort == true) { return 0; }
+
+ cout.setf(ios::fixed, ios::floatfield);
+ cout.setf(ios::showpoint);
+
+ getSingleLookUp();
+ getJointLookUp();
+
+
+ vector<string> flowFileVector;
+ if(flowFilesFileName != "not found"){
+ string fName;
+
+ ifstream flowFilesFile;
+ m->openInputFile(flowFilesFileName, flowFilesFile);
+ while(flowFilesFile){
+ flowFilesFile >> fName;
+ flowFileVector.push_back(fName);
+ m->gobble(flowFilesFile);
+ }
+ }
+ else{
+ flowFileVector.push_back(flowFileName);
+ }
+ int numFiles = flowFileVector.size();
+
+
+ for(int i=0;i<numFiles;i++){
+ flowFileName = flowFileVector[i];
+
+ cout << "\n>>>>>\tProcessing " << flowFileName << " (file " << i+1 << " of " << numFiles << ")\t<<<<<" << endl;
+ cout << "Reading flowgrams..." << endl;
+ getFlowData();
+ cout << "Identifying unique flowgrams..." << endl;
+ getUniques();
+
+
+ cout << "Calculating distances between flowgrams..." << endl;
+ string distFileName = createDistFile(processors);
+ string namesFileName = createNamesFile();
+
+ cout << "\nClustering flowgrams..." << endl;
+ string listFileName = cluster(distFileName, namesFileName);
+ getOTUData(listFileName);
+
+ initPyroCluster();
+
+ double maxDelta = 0;
+ int iter = 0;
+
+ double begClock = clock();
+ unsigned long int begTime = time(NULL);
+
+ cout << "\nDenoising flowgrams..." << endl;
+ cout << "iter\tmaxDelta\tnLL\t\tcycletime" << endl;
+
+ while((maxIters == 0 && maxDelta > minDelta) || iter < MIN_ITER || (maxDelta > minDelta && iter < maxIters)){
+
+ double cycClock = clock();
+ unsigned long int cycTime = time(NULL);
+ fill();
+
+ calcCentroids();
+
+ maxDelta = getNewWeights();
+ double nLL = getLikelihood();
+ checkCentroids();
+
+ calcNewDistances();
+
+ iter++;
+
+ cout << iter << '\t' << maxDelta << '\t' << setprecision(2) << nLL << '\t' << time(NULL) - cycTime << '\t' << setprecision(6) << (clock() - cycClock)/(double)CLOCKS_PER_SEC << endl;
+ }
+
+ cout << "\nFinalizing..." << endl;
+ fill();
+ setOTUs();
+
+ vector<int> otuCounts(numOTUs, 0);
+ for(int i=0;i<numSeqs;i++) { otuCounts[otuData[i]]++; }
+
+ calcCentroidsDriver(0, numOTUs);
+ writeQualities(otuCounts);
+ writeSequences(otuCounts);
+ writeNames(otuCounts);
+ writeClusters(otuCounts);
+ writeGroups();
+
+ remove(distFileName.c_str());
+ remove(namesFileName.c_str());
+ remove(listFileName.c_str());
+
+ cout << "Total time to process " << flowFileName << ":\t" << time(NULL) - begTime << '\t' << setprecision(6) << (clock() - begClock)/(double)CLOCKS_PER_SEC << endl;
+ }
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "execute");
+ exit(1);
+ }
+}
+#endif
+/**************************************************************************************************/
+
+void ShhherCommand::getFlowData(){
+ try{
+ ifstream flowFile;
+ m->openInputFile(flowFileName, flowFile);
+
+ string seqName;
+
+ int currentNumFlowCells;
+
+ float intensity;
+
+ flowFile >> numFlowCells;
+ int index = 0;//pcluster
+ while(!flowFile.eof()){
+ flowFile >> seqName >> currentNumFlowCells;
+ lengths.push_back(currentNumFlowCells);
+
+ seqNameVector.push_back(seqName);
+ nameMap[seqName] = index++;//pcluster
+
+ for(int i=0;i<numFlowCells;i++){
+ flowFile >> intensity;
+ if(intensity > 9.99) { intensity = 9.99; }
+ int intI = int(100 * intensity + 0.0001);
+ flowDataIntI.push_back(intI);
+ }
+ m->gobble(flowFile);
+ }
+ flowFile.close();
+
+ numSeqs = seqNameVector.size();
+
+ for(int i=0;i<numSeqs;i++){
+ int iNumFlowCells = i * numFlowCells;
+ for(int j=lengths[i];j<numFlowCells;j++){
+ flowDataIntI[iNumFlowCells + j] = 0;
+ }
+ }
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getFlowData");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::getSingleLookUp(){
+ try{
+ // these are the -log probabilities that a signal corresponds to a particular homopolymer length
+ singleLookUp.assign(HOMOPS * NUMBINS, 0);
+
+ int index = 0;
+ ifstream lookUpFile;
+ m->openInputFile(lookupFileName, lookUpFile);
+
+ for(int i=0;i<HOMOPS;i++){
+ float logFracFreq;
+ lookUpFile >> logFracFreq;
+
+ for(int j=0;j<NUMBINS;j++) {
+ lookUpFile >> singleLookUp[index];
+ index++;
+ }
+ }
+ lookUpFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getSingleLookUp");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::getJointLookUp(){
+ try{
+
+ // the most likely joint probability (-log) that two intenities have the same polymer length
+ jointLookUp.resize(NUMBINS * NUMBINS, 0);
+
+ for(int i=0;i<NUMBINS;i++){
+ for(int j=0;j<NUMBINS;j++){
+
+ float minSum = 10000000000;
+
+ for(int k=0;k<HOMOPS;k++){
+ float sum = singleLookUp[k * NUMBINS + i] + singleLookUp[k * NUMBINS + j];
+
+ if(sum < minSum) { minSum = sum; }
+ }
+ jointLookUp[i * NUMBINS + j] = minSum;
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getJointLookUp");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+float ShhherCommand::getProbIntensity(int intIntensity){
+ try{
+ float minNegLogProb = 10000000000;
+
+ for(int i=0;i<HOMOPS;i++){//loop signal strength
+ float negLogProb = singleLookUp[i * NUMBINS + intIntensity];
+ if(negLogProb < minNegLogProb) { minNegLogProb = negLogProb; }
+ }
+
+ return minNegLogProb;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getProbIntensity");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::getUniques(){
+ try{
+
+
+ numUniques = 0;
+ uniqueFlowgrams.assign(numFlowCells * numSeqs, -1);
+ uniqueCount.assign(numSeqs, 0); // anWeights
+ uniqueLengths.assign(numSeqs, 0);
+ mapSeqToUnique.assign(numSeqs, -1);
+ mapUniqueToSeq.assign(numSeqs, -1);
+
+ vector<short> uniqueFlowDataIntI(numFlowCells * numSeqs, -1);
+
+ for(int i=0;i<numSeqs;i++){
+ int index = 0;
+
+ vector<short> current(numFlowCells);
+ for(int j=0;j<numFlowCells;j++){ current[j] = short(((flowDataIntI[i * numFlowCells + j] + 50.0)/100.0)); }
+
+ for(int j=0;j<numUniques;j++){
+ int offset = j * numFlowCells;
+ bool toEnd = 1;
+
+ for(int k=0;k<numFlowCells;k++){
+ if(current[k] != uniqueFlowgrams[offset + k]){
+ toEnd = 0;
+ break;
+ }
+ }
+
+ if(toEnd){
+ mapSeqToUnique[i] = j;
+ uniqueCount[j]++;
+ index = j;
+ break;
+ }
+ index++;
+ }
+
+ if(index == numUniques){
+ uniqueLengths[numUniques] = lengths[i];
+ uniqueCount[numUniques] = 1;
+ mapSeqToUnique[i] = numUniques;//anMap
+ mapUniqueToSeq[numUniques] = i;//anF
+
+ for(int k=0;k<numFlowCells;k++){
+ uniqueFlowgrams[numUniques * numFlowCells + k] = current[k];
+ uniqueFlowDataIntI[numUniques * numFlowCells + k] = flowDataIntI[i * numFlowCells + k];
+ }
+
+ numUniques++;
+ }
+ }
+ uniqueFlowDataIntI.resize(numFlowCells * numUniques);
+ uniqueLengths.resize(numUniques);
+
+ flowDataPrI.assign(numSeqs * numFlowCells, 0);
+ for(int i=0;i<flowDataPrI.size();i++) { flowDataPrI[i] = getProbIntensity(flowDataIntI[i]); }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getUniques");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+float ShhherCommand::calcPairwiseDist(int seqA, int seqB){
+ try{
+ int minLength = lengths[mapSeqToUnique[seqA]];
+ if(lengths[seqB] < minLength){ minLength = lengths[mapSeqToUnique[seqB]]; }
+
+ int ANumFlowCells = seqA * numFlowCells;
+ int BNumFlowCells = seqB * numFlowCells;
+
+ float dist = 0;
+
+ for(int i=0;i<minLength;i++){
+ int flowAIntI = flowDataIntI[ANumFlowCells + i];
+ float flowAPrI = flowDataPrI[ANumFlowCells + i];
+
+ int flowBIntI = flowDataIntI[BNumFlowCells + i];
+ float flowBPrI = flowDataPrI[BNumFlowCells + i];
+ dist += jointLookUp[flowAIntI * NUMBINS + flowBIntI] - flowAPrI - flowBPrI;
+ }
+
+ dist /= (float) minLength;
+ return dist;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcPairwiseDist");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::flowDistParentFork(string distFileName, int startSeq, int stopSeq){
+ try{
+
+ ostringstream outStream;
+ outStream.setf(ios::fixed, ios::floatfield);
+ outStream.setf(ios::dec, ios::basefield);
+ outStream.setf(ios::showpoint);
+ outStream.precision(6);
+
+ int begTime = time(NULL);
+ double begClock = clock();
+
+ for(int i=startSeq;i<stopSeq;i++){
+ for(int j=0;j<i;j++){
+ float flowDistance = calcPairwiseDist(mapUniqueToSeq[i], mapUniqueToSeq[j]);
+
+ if(flowDistance < 1e-6){
+ outStream << seqNameVector[mapUniqueToSeq[i]] << '\t' << seqNameVector[mapUniqueToSeq[j]] << '\t' << 0.000000 << endl;
+ }
+ else if(flowDistance <= cutoff){
+ outStream << seqNameVector[mapUniqueToSeq[i]] << '\t' << seqNameVector[mapUniqueToSeq[j]] << '\t' << flowDistance << endl;
+ }
+ }
+ if(i % 100 == 0){
+ m->mothurOut(toString(i) + "\t" + toString(time(NULL) - begTime));
+ m->mothurOut("\t" + toString((clock()-begClock)/CLOCKS_PER_SEC));
+ m->mothurOutEndLine();
+ }
+ }
+ m->mothurOut(toString(stopSeq-1) + "\t" + toString(time(NULL) - begTime));
+ m->mothurOut("\t" + toString((clock()-begClock)/CLOCKS_PER_SEC));
+ m->mothurOutEndLine();
+
+ ofstream distFile((distFileName + "temp." + toString(pid)).c_str());
+ distFile << outStream.str();
+ distFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "flowDistParentFork");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+string ShhherCommand::createDistFile(int processors){
+ try{
+ string fDistFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.dist";
+
+ unsigned long int begTime = time(NULL);
+ double begClock = clock();
+
+ vector<int> start;
+ vector<int> end;
+
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ if(processors == 1) { flowDistParentFork(fDistFileName, 0, numUniques); }
+ else{ //you have multiple processors
+
+ if (numSeqs < processors){ processors = 1; }
+
+ vector<int> start(processors, 0);
+ vector<int> end(processors, 0);
+
+ for (int i = 0; i < processors; i++) {
+ start[i] = int(sqrt(float(i)/float(processors)) * numUniques);
+ end[i] = int(sqrt(float(i+1)/float(processors)) * numUniques);
+ }
+
+ int process = 1;
+ vector<int> processIDs;
+
+ //loop through and create all the processes you want
+ while (process != processors) {
+ int pid = fork();
+
+ if (pid > 0) {
+ processIDs.push_back(pid); //create map from line number to pid so you can append files in correct order later
+ process++;
+ }else if (pid == 0){
+ flowDistParentFork(fDistFileName + toString(getpid()) + ".temp", start[process], end[process]);
+ exit(0);
+ }else {
+ m->mothurOut("[ERROR]: unable to spawn the necessary processes. Error code: " + toString(pid)); m->mothurOutEndLine();
+ perror(" : ");
+ for (int i=0;i<processIDs.size();i++) { int temp = processIDs[i]; kill (temp, SIGINT); }
+ exit(0);
+ }
+ }
+
+ //parent does its part
+ flowDistParentFork(fDistFileName, start[0], end[0]);
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processIDs.size();i++) {
+ int temp = processIDs[i];
+ wait(&temp);
+ }
+
+ //append and remove temp files
+ for (int i=0;i<processIDs.size();i++) {
+ m->appendFiles((fDistFileName + toString(processIDs[i]) + ".temp"), fDistFileName);
+ remove((fDistFileName + toString(processIDs[i]) + ".temp").c_str());
+ }
+
+ }
+
+#else
+ flowDistParentFork(fDistFileName, 0, numUniques);
+#endif
+
+ m->mothurOutEndLine();
+
+ cout << "Total time: " << (time(NULL) - begTime) << "\t" << (clock() - begClock)/CLOCKS_PER_SEC << endl;;
+
+ return fDistFileName;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "createDistFile");
+ exit(1);
+ }
+
+}
+
+/**************************************************************************************************/
+
+string ShhherCommand::createNamesFile(){
+ try{
+
+ vector<string> duplicateNames(numUniques, "");
+ for(int i=0;i<numSeqs;i++){
+ duplicateNames[mapSeqToUnique[i]] += seqNameVector[i] + ',';
+ }
+
+ string nameFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.names";
+
+ ofstream nameFile;
+ m->openOutputFile(nameFileName, nameFile);
+
+ for(int i=0;i<numUniques;i++){
+// nameFile << seqNameVector[mapUniqueToSeq[i]] << '\t' << duplicateNames[i].substr(0, duplicateNames[i].find_last_of(',')) << endl;
+ nameFile << mapUniqueToSeq[i] << '\t' << duplicateNames[i].substr(0, duplicateNames[i].find_last_of(',')) << endl;
+ }
+
+ nameFile.close();
+ return nameFileName;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "createNamesFile");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+string ShhherCommand::cluster(string distFileName, string namesFileName){
+ try {
+
+ SparseMatrix* matrix;
+ ListVector* list;
+ RAbundVector* rabund;
+
+ globaldata->setNameFile(namesFileName);
+ globaldata->setColumnFile(distFileName);
+ globaldata->setFormat("column");
+
+ ReadMatrix* read = new ReadColumnMatrix(distFileName);
+ read->setCutoff(cutoff);
+
+ NameAssignment* clusterNameMap = new NameAssignment(namesFileName);
+ clusterNameMap->readMap();
+ read->read(clusterNameMap);
+
+ list = read->getListVector();
+ matrix = read->getMatrix();
+
+ delete read;
+ delete clusterNameMap;
+
+ rabund = new RAbundVector(list->getRAbundVector());
+
+ Cluster* cluster = new CompleteLinkage(rabund, list, matrix, cutoff, "furthest");
+ string tag = cluster->getTag();
+
+ double clusterCutoff = cutoff;
+ while (matrix->getSmallDist() <= clusterCutoff && matrix->getNNodes() > 0){
+ cluster->update(clusterCutoff);
+ float dist = matrix->getSmallDist();
+ }
+
+ list->setLabel(toString(cutoff));
+
+ string listFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.list";
+ ofstream listFile;
+ m->openOutputFile(listFileName, listFile);
+ list->print(listFile);
+ listFile.close();
+
+ delete matrix; delete cluster; delete rabund; delete list;
+
+ return listFileName;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "cluster");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::getOTUData(string listFileName){
+ try {
+
+ ifstream listFile;
+ m->openInputFile(listFileName, listFile);
+ string label;
+
+ listFile >> label >> numOTUs;
+
+ otuData.assign(numSeqs, 0);
+ cumNumSeqs.assign(numOTUs, 0);
+ nSeqsPerOTU.assign(numOTUs, 0);
+ aaP.resize(numOTUs);
+
+ string singleOTU = "";
+
+ for(int i=0;i<numOTUs;i++){
+
+ listFile >> singleOTU;
+
+ istringstream otuString(singleOTU);
+
+ while(otuString){
+
+ string seqName = "";
+
+ for(int j=0;j<singleOTU.length();j++){
+ char letter = otuString.get();
+
+ if(letter != ','){
+ seqName += letter;
+ }
+ else{
+ map<string,int>::iterator nmIt = nameMap.find(seqName);
+ int index = nmIt->second;
+
+ nameMap.erase(nmIt);
+
+ otuData[index] = i;
+ nSeqsPerOTU[i]++;
+ aaP[i].push_back(index);
+ seqName = "";
+ }
+ }
+
+ map<string,int>::iterator nmIt = nameMap.find(seqName);
+
+ int index = nmIt->second;
+ nameMap.erase(nmIt);
+
+ otuData[index] = i;
+ nSeqsPerOTU[i]++;
+ aaP[i].push_back(index);
+
+ otuString.get();
+ }
+
+ sort(aaP[i].begin(), aaP[i].end());
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ seqNumber.push_back(aaP[i][j]);
+ }
+ for(int j=nSeqsPerOTU[i];j<numSeqs;j++){
+ aaP[i].push_back(0);
+ }
+ }
+
+ for(int i=1;i<numOTUs;i++){
+ cumNumSeqs[i] = cumNumSeqs[i-1] + nSeqsPerOTU[i-1];
+ }
+ aaI = aaP;
+ seqIndex = seqNumber;
+
+ listFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getOTUData");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::initPyroCluster(){
+ try{
+ dist.assign(numSeqs * numOTUs, 0);
+ change.assign(numOTUs, 1);
+ centroids.assign(numOTUs, -1);
+ weight.assign(numOTUs, 0);
+ singleTau.assign(numSeqs, 1.0);
+
+ nSeqsBreaks.assign(processors+1, 0);
+ nOTUsBreaks.assign(processors+1, 0);
+
+ nSeqsBreaks[0] = 0;
+ for(int i=0;i<processors;i++){
+ nSeqsBreaks[i+1] = nSeqsBreaks[i] + (int)((double) numSeqs / (double) processors);
+ nOTUsBreaks[i+1] = nOTUsBreaks[i] + (int)((double) numOTUs / (double) processors);
+ }
+ nSeqsBreaks[processors] = numSeqs;
+ nOTUsBreaks[processors] = numOTUs;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "initPyroCluster");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::fill(){
+ try {
+ int index = 0;
+ for(int i=0;i<numOTUs;i++){
+ cumNumSeqs[i] = index;
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ seqNumber[index] = aaP[i][j];
+ seqIndex[index] = aaI[i][j];
+
+ index++;
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "fill");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::calcCentroids(){
+ try{
+
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+
+ if(processors == 1) {
+ calcCentroidsDriver(0, numOTUs);
+ }
+ else{ //you have multiple processors
+ if (numOTUs < processors){ processors = 1; }
+
+ int process = 1;
+ vector<int> processIDs;
+
+ //loop through and create all the processes you want
+ while (process != processors) {
+ int pid = vfork();
+
+ if (pid > 0) {
+ processIDs.push_back(pid); //create map from line number to pid so you can append files in correct order later
+ process++;
+ }else if (pid == 0){
+ calcCentroidsDriver(nOTUsBreaks[process], nOTUsBreaks[process+1]);
+ exit(0);
+ }else {
+ m->mothurOut("[ERROR]: unable to spawn the necessary processes. Error code: " + toString(pid)); m->mothurOutEndLine();
+ perror(" : ");
+ for (int i=0;i<processIDs.size();i++) { int temp = processIDs[i]; kill (temp, SIGINT); }
+ exit(0);
+ }
+ }
+
+ //parent does its part
+ calcCentroidsDriver(nOTUsBreaks[0], nOTUsBreaks[1]);
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processIDs.size();i++) {
+ int temp = processIDs[i];
+ wait(&temp);
+ }
+ }
+
+#else
+ calcCentroidsDriver(0, numOTUs);
+#endif
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcCentroidsDriver");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::calcCentroidsDriver(int start, int finish){
+
+ //this function gets the most likely homopolymer length at a flow position for a group of sequences
+ //within an otu
+
+ try{
+
+ for(int i=start;i<finish;i++){
+
+ double count = 0;
+ int position = 0;
+ int minFlowGram = 100000000;
+ double minFlowValue = 1e8;
+ change[i] = 0; //FALSE
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ count += singleTau[seqNumber[cumNumSeqs[i] + j]];
+ }
+
+ if(nSeqsPerOTU[i] > 0 && count > MIN_COUNT){
+ vector<double> adF(nSeqsPerOTU[i]);
+ vector<int> anL(nSeqsPerOTU[i]);
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int index = cumNumSeqs[i] + j;
+ int nI = seqIndex[index];
+ int nIU = mapSeqToUnique[nI];
+
+ int k;
+ for(k=0;k<position;k++){
+ if(nIU == anL[k]){
+ break;
+ }
+ }
+ if(k == position){
+ anL[position] = nIU;
+ adF[position] = 0.0000;
+ position++;
+ }
+ }
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int index = cumNumSeqs[i] + j;
+ int nI = seqIndex[index];
+ int nIU = mapSeqToUnique[nI];
+
+ double tauValue = singleTau[seqNumber[index]];
+
+ for(int k=0;k<position;k++){
+ double dist = getDistToCentroid(anL[k], nI, lengths[nI]);
+ adF[k] += dist * tauValue;
+ }
+ }
+
+ for(int j=0;j<position;j++){
+ if(adF[j] < minFlowValue){
+ minFlowGram = j;
+ minFlowValue = adF[j];
+ }
+ }
+
+ if(centroids[i] != anL[minFlowGram]){
+ change[i] = 1;
+ centroids[i] = anL[minFlowGram];
+ }
+ }
+ else if(centroids[i] != -1){
+ change[i] = 1;
+ centroids[i] = -1;
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcCentroidsDriver");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+double ShhherCommand::getDistToCentroid(int cent, int flow, int length){
+ try{
+
+ int flowAValue = cent * numFlowCells;
+ int flowBValue = flow * numFlowCells;
+
+ double dist = 0;
+
+ for(int i=0;i<length;i++){
+ dist += singleLookUp[uniqueFlowgrams[flowAValue] * NUMBINS + flowDataIntI[flowBValue]];
+ flowAValue++;
+ flowBValue++;
+ }
+
+ return dist / (double)length;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getDistToCentroid");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+double ShhherCommand::getNewWeights(){
+ try{
+
+ double maxChange = 0;
+
+ for(int i=0;i<numOTUs;i++){
+
+ double difference = weight[i];
+ weight[i] = 0;
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int index = cumNumSeqs[i] + j;
+ int nI = seqIndex[index];
+ double tauValue = singleTau[seqNumber[index]];
+ weight[i] += tauValue;
+ }
+
+ difference = fabs(weight[i] - difference);
+ if(difference > maxChange){ maxChange = difference; }
+ }
+ return maxChange;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getNewWeights");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+double ShhherCommand::getLikelihood(){
+
+ try{
+
+ vector<double> P(numSeqs, 0);
+ int effNumOTUs = 0;
+
+ for(int i=0;i<numOTUs;i++){
+ if(weight[i] > MIN_WEIGHT){
+ effNumOTUs++;
+ }
+ }
+
+ for(int i=0;i<numOTUs;i++){
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int index = cumNumSeqs[i] + j;
+ int nI = seqIndex[index];
+ double singleDist = dist[seqNumber[index]];
+
+ P[nI] += weight[i] * exp(-singleDist * sigma);
+ }
+ }
+
+ double nLL = 0.00;
+ for(int i=0;i<numSeqs;i++){
+ nLL += -log(P[i]);
+ }
+
+ nLL = nLL -(double)numSeqs * log(sigma);
+
+ return nLL;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "getNewWeights");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::checkCentroids(){
+ try{
+ vector<int> unique(numOTUs, 1);
+
+ for(int i=0;i<numOTUs;i++){
+ if(centroids[i] == -1 || weight[i] < MIN_WEIGHT){
+ unique[i] = -1;
+ }
+ }
+
+ for(int i=0;i<numOTUs;i++){
+ if(unique[i] == 1){
+ for(int j=i+1;j<numOTUs;j++){
+ if(unique[j] == 1){
+
+ if(centroids[j] == centroids[i]){
+ unique[j] = 0;
+ centroids[j] = -1;
+
+ weight[i] += weight[j];
+ weight[j] = 0.0;
+ }
+ }
+ }
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "checkCentroids");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::calcNewDistances(){
+ try{
+
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+
+ if(processors == 1) {
+ calcNewDistancesParent(0, numSeqs);
+ }
+ else{ //you have multiple processors
+ if (numSeqs < processors){ processors = 1; }
+
+ vector<vector<int> > child_otuIndex(processors);
+ vector<vector<int> > child_seqIndex(processors);
+ vector<vector<double> > child_singleTau(processors);
+ vector<int> totals(processors);
+
+ int process = 1;
+ vector<int> processIDs;
+
+ //loop through and create all the processes you want
+ while (process != processors) {
+ int pid = vfork();
+
+ if (pid > 0) {
+ processIDs.push_back(pid); //create map from line number to pid so you can append files in correct order later
+ process++;
+ }else if (pid == 0){
+ calcNewDistancesChild(nSeqsBreaks[process], nSeqsBreaks[process+1], child_otuIndex[process], child_seqIndex[process], child_singleTau[process]);
+ totals[process] = child_otuIndex[process].size();
+
+ exit(0);
+ }else {
+ m->mothurOut("[ERROR]: unable to spawn the necessary processes. Error code: " + toString(pid)); m->mothurOutEndLine();
+ perror(" : ");
+ for (int i=0;i<processIDs.size();i++) { int temp = processIDs[i]; kill (temp, SIGINT); }
+ exit(0);
+ }
+ }
+
+ //parent does its part
+ calcNewDistancesParent(nSeqsBreaks[0], nSeqsBreaks[1]);
+ int total = seqIndex.size();
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processIDs.size();i++) {
+ int temp = processIDs[i];
+ wait(&temp);
+ }
+
+ for(int i=1;i<processors;i++){
+ int oldTotal = total;
+ total += totals[i];
+
+ singleTau.resize(total, 0);
+ seqIndex.resize(total, 0);
+ seqNumber.resize(total, 0);
+
+ int childIndex = 0;
+
+ for(int j=oldTotal;j<total;j++){
+ int otuI = child_otuIndex[i][childIndex];
+ int seqI = child_seqIndex[i][childIndex];
+
+ singleTau[j] = child_singleTau[i][childIndex];
+ aaP[otuI][nSeqsPerOTU[otuI]] = j;
+ aaI[otuI][nSeqsPerOTU[otuI]] = seqI;
+ nSeqsPerOTU[otuI]++;
+
+ childIndex++;
+ }
+ }
+ }
+#else
+ calcNewDistancesParent(0, numSeqs);
+#endif
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcNewDistances");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+#ifdef USE_MPI
+void ShhherCommand::calcNewDistancesChildMPI(int startSeq, int stopSeq, vector<int>& otuIndex){
+
+ try{
+ vector<double> newTau(numOTUs,0);
+ vector<double> norms(numSeqs, 0);
+ otuIndex.resize(0);
+ seqIndex.resize(0);
+ singleTau.resize(0);
+
+
+
+ for(int i=startSeq;i<stopSeq;i++){
+ double offset = 1e8;
+ int indexOffset = i * numOTUs;
+
+ for(int j=0;j<numOTUs;j++){
+
+ if(weight[j] > MIN_WEIGHT && change[j] == 1){
+ dist[indexOffset + j] = getDistToCentroid(centroids[j], i, lengths[i]);
+ }
+ if(weight[j] > MIN_WEIGHT && dist[indexOffset + j] < offset){
+ offset = dist[indexOffset + j];
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ if(weight[j] > MIN_WEIGHT){
+ newTau[j] = exp(sigma * (-dist[indexOffset + j] + offset)) * weight[j];
+ norms[i] += newTau[j];
+ }
+ else{
+ newTau[j] = 0.0;
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+
+ newTau[j] /= norms[i];
+
+ if(newTau[j] > MIN_TAU){
+ otuIndex.push_back(j);
+ seqIndex.push_back(i);
+ singleTau.push_back(newTau[j]);
+ }
+ }
+
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcNewDistancesChildMPI");
+ exit(1);
+ }
+}
+#endif
+/**************************************************************************************************/
+
+void ShhherCommand::calcNewDistancesChild(int startSeq, int stopSeq, vector<int>& child_otuIndex, vector<int>& child_seqIndex, vector<double>& child_singleTau){
+
+ try{
+ vector<double> newTau(numOTUs,0);
+ vector<double> norms(numSeqs, 0);
+ child_otuIndex.resize(0);
+ child_seqIndex.resize(0);
+ child_singleTau.resize(0);
+
+ for(int i=startSeq;i<stopSeq;i++){
+ double offset = 1e8;
+ int indexOffset = i * numOTUs;
+
+
+ for(int j=0;j<numOTUs;j++){
+ if(weight[j] > MIN_WEIGHT && change[j] == 1){
+ dist[indexOffset + j] = getDistToCentroid(centroids[j], i, lengths[i]);
+ }
+
+ if(weight[j] > MIN_WEIGHT && dist[indexOffset + j] < offset){
+ offset = dist[indexOffset + j];
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ if(weight[j] > MIN_WEIGHT){
+ newTau[j] = exp(sigma * (-dist[indexOffset + j] + offset)) * weight[j];
+ norms[i] += newTau[j];
+ }
+ else{
+ newTau[j] = 0.0;
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ newTau[j] /= norms[i];
+
+ if(newTau[j] > MIN_TAU){
+ child_otuIndex.push_back(j);
+ child_seqIndex.push_back(i);
+ child_singleTau.push_back(newTau[j]);
+ }
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcNewDistancesChild");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::calcNewDistancesParent(int startSeq, int stopSeq){
+
+ try{
+
+ int total = 0;
+ vector<double> newTau(numOTUs,0);
+ vector<double> norms(numSeqs, 0);
+ nSeqsPerOTU.assign(numOTUs, 0);
+
+ for(int i=startSeq;i<stopSeq;i++){
+ int indexOffset = i * numOTUs;
+
+ double offset = 1e8;
+
+ for(int j=0;j<numOTUs;j++){
+ if(weight[j] > MIN_WEIGHT && change[j] == 1){
+ dist[indexOffset + j] = getDistToCentroid(centroids[j], i, lengths[i]);
+ }
+
+ if(weight[j] > MIN_WEIGHT && dist[indexOffset + j] < offset){
+ offset = dist[indexOffset + j];
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ if(weight[j] > MIN_WEIGHT){
+ newTau[j] = exp(sigma * (-dist[indexOffset + j] + offset)) * weight[j];
+ norms[i] += newTau[j];
+ }
+ else{
+ newTau[j] = 0.0;
+ }
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ newTau[j] /= norms[i];
+ }
+
+ for(int j=0;j<numOTUs;j++){
+ if(newTau[j] > MIN_TAU){
+
+ int oldTotal = total;
+
+ total++;
+
+ singleTau.resize(total, 0);
+ seqNumber.resize(total, 0);
+ seqIndex.resize(total, 0);
+
+ singleTau[oldTotal] = newTau[j];
+
+ aaP[j][nSeqsPerOTU[j]] = oldTotal;
+ aaI[j][nSeqsPerOTU[j]] = i;
+ nSeqsPerOTU[j]++;
+ }
+ }
+ }
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcNewDistancesParent");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::setOTUs(){
+
+ try {
+ vector<double> bigTauMatrix(numOTUs * numSeqs, 0.0000);
+
+ for(int i=0;i<numOTUs;i++){
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int index = cumNumSeqs[i] + j;
+ double tauValue = singleTau[seqNumber[index]];
+ int sIndex = seqIndex[index];
+ bigTauMatrix[sIndex * numOTUs + i] = tauValue;
+ }
+ }
+
+ for(int i=0;i<numSeqs;i++){
+ double maxTau = -1.0000;
+ int maxOTU = -1;
+ for(int j=0;j<numOTUs;j++){
+ if(bigTauMatrix[i * numOTUs + j] > maxTau){
+ maxTau = bigTauMatrix[i * numOTUs + j];
+ maxOTU = j;
+ }
+ }
+
+ otuData[i] = maxOTU;
+ }
+
+ nSeqsPerOTU.assign(numOTUs, 0);
+
+ for(int i=0;i<numSeqs;i++){
+ int index = otuData[i];
+
+ singleTau[i] = 1.0000;
+ dist[i] = 0.0000;
+
+ aaP[index][nSeqsPerOTU[index]] = i;
+ aaI[index][nSeqsPerOTU[index]] = i;
+
+ nSeqsPerOTU[index]++;
+ }
+ fill();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "calcNewDistances");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::writeQualities(vector<int> otuCounts){
+
+ try {
+ string qualityFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.qual";
+
+ ofstream qualityFile;
+ m->openOutputFile(qualityFileName, qualityFile);
+
+ qualityFile.setf(ios::fixed, ios::floatfield);
+ qualityFile.setf(ios::showpoint);
+ qualityFile << setprecision(6);
+
+ vector<vector<int> > qualities(numOTUs);
+ vector<double> pr(HOMOPS, 0);
+
+ int index = 0;
+
+ for(int i=0;i<numOTUs;i++){
+ int index = 0;
+ int base = 0;
+
+ if(nSeqsPerOTU[i] > 0){
+ qualities[i].assign(1024, -1);
+
+ while(index < numFlowCells){
+ double maxPrValue = 1e8;
+ short maxPrIndex = -1;
+ double count = 0.0000;
+
+ pr.assign(HOMOPS, 0);
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int lIndex = cumNumSeqs[i] + j;
+ double tauValue = singleTau[seqNumber[lIndex]];
+ int sequenceIndex = aaI[i][j];
+ short intensity = flowDataIntI[sequenceIndex * numFlowCells + index];
+
+ count += tauValue;
+
+ for(int s=0;s<HOMOPS;s++){
+ pr[s] += tauValue * singleLookUp[s * NUMBINS + intensity];
+ }
+ }
+
+ maxPrIndex = uniqueFlowgrams[centroids[i] * numFlowCells + index];
+ maxPrValue = pr[maxPrIndex];
+
+ if(count > MIN_COUNT){
+ double U = 0.0000;
+ double norm = 0.0000;
+
+ for(int s=0;s<HOMOPS;s++){
+ norm += exp(-(pr[s] - maxPrValue));
+ }
+
+ for(int s=1;s<=maxPrIndex;s++){
+ int value = 0;
+ double temp = 0.0000;
+
+ U += exp(-(pr[s-1]-maxPrValue))/norm;
+
+ if(U>0.00){
+ temp = log10(U);
+ }
+ else{
+ temp = -10.1;
+ }
+ temp = floor(-10 * temp);
+ value = (int)floor(temp);
+ if(value > 100){ value = 100; }
+
+ qualities[i][base] = (int)value;
+ base++;
+ }
+ }
+
+ index++;
+ }
+ }
+
+
+ if(otuCounts[i] > 0){
+ qualityFile << '>' << seqNameVector[mapUniqueToSeq[i]] << endl;
+
+ int j=4; //need to get past the first four bases
+ while(qualities[i][j] != -1){
+ qualityFile << qualities[i][j] << ' ';
+ j++;
+ }
+ qualityFile << endl;
+ }
+ }
+ qualityFile.close();
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "writeQualities");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::writeSequences(vector<int> otuCounts){
+ try {
+
+ string bases = "TACG";
+
+ string fastaFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.fasta";
+ ofstream fastaFile;
+ m->openOutputFile(fastaFileName, fastaFile);
+
+ vector<string> names(numOTUs, "");
+
+ for(int i=0;i<numOTUs;i++){
+ int index = centroids[i];
+
+ if(otuCounts[i] > 0){
+ fastaFile << '>' << seqNameVector[aaI[i][0]] << endl;
+
+ for(int j=8;j<numFlowCells;j++){
+
+ char base = bases[j % 4];
+ for(int k=0;k<uniqueFlowgrams[index * numFlowCells + j];k++){
+ fastaFile << base;
+ }
+ }
+ fastaFile << endl;
+ }
+ }
+ fastaFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "writeSequences");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::writeNames(vector<int> otuCounts){
+ try {
+ string nameFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.final.names";
+ ofstream nameFile;
+ m->openOutputFile(nameFileName, nameFile);
+
+ for(int i=0;i<numOTUs;i++){
+ if(otuCounts[i] > 0){
+ nameFile << seqNameVector[aaI[i][0]] << '\t' << seqNameVector[aaI[i][0]];
+
+ for(int j=1;j<nSeqsPerOTU[i];j++){
+ nameFile << ',' << seqNameVector[aaI[i][j]];
+ }
+
+ nameFile << endl;
+ }
+ }
+ nameFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "writeNames");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::writeGroups(){
+ try {
+ string fileRoot = flowFileName.substr(0,flowFileName.find_last_of('.'));
+ string groupFileName = fileRoot + ".pn.groups";
+ ofstream groupFile;
+ m->openOutputFile(groupFileName, groupFile);
+
+ for(int i=0;i<numSeqs;i++){
+ groupFile << seqNameVector[i] << '\t' << fileRoot << endl;
+ }
+ groupFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "writeGroups");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+void ShhherCommand::writeClusters(vector<int> otuCounts){
+ try {
+ string otuCountsFileName = flowFileName.substr(0,flowFileName.find_last_of('.')) + ".pn.counts";
+ ofstream otuCountsFile;
+ m->openOutputFile(otuCountsFileName, otuCountsFile);
+
+ string bases = "TACG";
+
+ for(int i=0;i<numOTUs;i++){
+ //output the translated version of the centroid sequence for the otu
+ if(otuCounts[i] > 0){
+ int index = centroids[i];
+
+ otuCountsFile << "ideal\t";
+ for(int j=8;j<numFlowCells;j++){
+ char base = bases[j % 4];
+ for(int s=0;s<uniqueFlowgrams[index * numFlowCells + j];s++){
+ otuCountsFile << base;
+ }
+ }
+ otuCountsFile << endl;
+
+ for(int j=0;j<nSeqsPerOTU[i];j++){
+ int sequence = aaI[i][j];
+ otuCountsFile << seqNameVector[sequence] << '\t';
+
+ for(int k=8;k<lengths[sequence];k++){
+ char base = bases[k % 4];
+ int freq = int(0.01 * (double)flowDataIntI[sequence * numFlowCells + k] + 0.5);
+
+ for(int s=0;s<freq;s++){
+ otuCountsFile << base;
+ }
+ }
+ otuCountsFile << endl;
+ }
+ otuCountsFile << endl;
+ }
+ }
+ otuCountsFile.close();
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ShhherCommand", "writeClusters");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************