+ m->errorOut(e, "DistanceCommand", "driver");
+ exit(1);
+ }
+}
+#ifdef USE_MPI
+/**************************************************************************************************/
+/////// need to fix to work with calcs and sequencedb
+int DistanceCommand::driverMPI(int startLine, int endLine, MPI_File& outMPI, float cutoff){
+ try {
+ MPI_Status status;
+ int startTime = time(NULL);
+
+ string outputString = "";
+
+ for(int i=startLine;i<endLine;i++){
+
+ for(int j=0;j<i;j++){
+
+ if (m->control_pressed) { return 0; }
+
+ //if there was a column file given and we are appending, we don't want to calculate the distances that are already in the column file
+ //the alignDB contains the new sequences and then the old, so if i an oldsequence and j is an old sequence then break out of this loop
+ if ((i >= numNewFasta) && (j >= numNewFasta)) { break; }
+
+ distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
+ double dist = distCalculator->getDist();
+
+ if(dist <= cutoff){
+ if (output == "column") { outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); }
+ }
+
+ if (output == "square") { //make a square column you can convert to square phylip
+ outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n');
+ outputString += (alignDB.get(j).getName() + ' ' + alignDB.get(i).getName() + ' ' + toString(dist) + '\n');
+ }
+ }
+
+ if(i % 100 == 0){
+ //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+ cout << i << '\t' << (time(NULL) - startTime) << endl;
+ }
+
+
+ //send results to parent
+ int length = outputString.length();
+
+ char* buf = new char[length];
+ memcpy(buf, outputString.c_str(), length);
+
+ MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status);
+ outputString = "";
+ delete buf;
+
+ }
+
+ //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+ cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
+ return 1;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "driverMPI");
+ exit(1);
+ }
+}
+/**************************************************************************************************/
+/////// need to fix to work with calcs and sequencedb
+int DistanceCommand::driverMPI(int startLine, int endLine, string file, long& size){
+ try {
+ MPI_Status status;
+
+ MPI_File outMPI;
+ int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
+
+ //char* filename = new char[file.length()];
+ //memcpy(filename, file.c_str(), file.length());
+
+ char filename[1024];
+ strcpy(filename, file.c_str());
+
+ MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
+ //delete filename;
+
+ int startTime = time(NULL);
+
+ string outputString = "";
+ size = 0;
+
+ if((output == "lt") && startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n"; }
+
+ for(int i=startLine;i<endLine;i++){
+ if(output == "lt") {
+ string name = alignDB.get(i).getName();
+ if (name.length() < 10) { //pad with spaces to make compatible
+ while (name.length() < 10) { name += " "; }
+ }
+ outputString += name + "\t";
+ }
+ for(int j=0;j<i;j++){
+
+ if (m->control_pressed) { return 0; }
+
+ distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
+ double dist = distCalculator->getDist();
+
+ if (output == "lt") { outputString += toString(dist) + "\t"; }
+ }
+
+ if (output == "lt") { outputString += "\n"; }
+
+
+ if(i % 100 == 0){
+ //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+ cout << i << '\t' << (time(NULL) - startTime) << endl;
+ }
+
+
+ //send results to parent
+ int length = outputString.length();
+ char* buf = new char[length];
+ memcpy(buf, outputString.c_str(), length);
+
+ MPI_File_write(outMPI, buf, length, MPI_CHAR, &status);
+ size += outputString.length();
+ outputString = "";
+ delete buf;
+ }
+
+ //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+ cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
+ MPI_File_close(&outMPI);
+
+ return 1;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "driverMPI");
+ exit(1);
+ }
+}
+#endif
+/**************************************************************************************************/
+int DistanceCommand::convertMatrix(string outputFile) {
+ try{
+
+ //sort file by first column so the distances for each row are together
+ string outfile = getRootName(outputFile) + "sorted.dist.temp";
+
+ //use the unix sort
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ string command = "sort -n " + outputFile + " -o " + outfile;
+ system(command.c_str());
+ #else //sort using windows sort
+ string command = "sort " + outputFile + " /O " + outfile;
+ system(command.c_str());
+ #endif
+
+
+ //output to new file distance for each row and save positions in file where new row begins
+ ifstream in;
+ openInputFile(outfile, in);
+
+ ofstream out;
+ openOutputFile(outputFile, out);
+
+ out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
+
+ out << alignDB.getNumSeqs() << endl;
+
+ //get first currentRow
+ string first, currentRow, second;
+ float dist;
+ map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
+ map<string, float>::iterator it;
+
+ in >> first;
+ currentRow = first;
+
+ rowDists[first] = 0.00; //distance to yourself is 0.0
+
+ in.seekg(0);
+ //openInputFile(outfile, in);
+
+ while(!in.eof()) {
+ if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
+
+ in >> first >> second >> dist; gobble(in);
+
+ if (first != currentRow) {
+ //print out last row
+ out << currentRow << '\t'; //print name
+
+ //print dists
+ for (it = rowDists.begin(); it != rowDists.end(); it++) {
+ out << it->second << '\t';
+ }
+ out << endl;
+
+ //start new row
+ currentRow = first;
+ rowDists.clear();
+ rowDists[first] = 0.00;
+ rowDists[second] = dist;
+ }else{
+ rowDists[second] = dist;
+ }
+ }
+ //print out last row
+ out << currentRow << '\t'; //print name
+
+ //print dists
+ for (it = rowDists.begin(); it != rowDists.end(); it++) {
+ out << it->second << '\t';
+ }
+ out << endl;
+
+ in.close();
+ out.close();
+
+ remove(outfile.c_str());
+
+ return 1;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "convertMatrix");
+ exit(1);
+ }
+}
+/**************************************************************************************************/
+int DistanceCommand::convertToLowerTriangle(string outputFile) {
+ try{
+
+ //sort file by first column so the distances for each row are together
+ string outfile = getRootName(outputFile) + "sorted.dist.temp";
+
+ //use the unix sort
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ string command = "sort -n " + outputFile + " -o " + outfile;
+ system(command.c_str());
+ #else //sort using windows sort
+ string command = "sort " + outputFile + " /O " + outfile;
+ system(command.c_str());
+ #endif
+
+
+ //output to new file distance for each row and save positions in file where new row begins
+ ifstream in;
+ openInputFile(outfile, in);
+
+ ofstream out;
+ openOutputFile(outputFile, out);
+
+ out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
+
+ out << alignDB.getNumSeqs() << endl;
+
+ //get first currentRow
+ string first, currentRow, second;
+ float dist;
+ int i, j;
+ i = 0; j = 0;
+ map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
+ map<string, float>::iterator it;
+
+ in >> first;
+ currentRow = first;
+
+ rowDists[first] = 0.00; //distance to yourself is 0.0
+
+ in.seekg(0);
+ //openInputFile(outfile, in);
+
+ while(!in.eof()) {
+ if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
+
+ in >> first >> second >> dist; gobble(in);
+
+ if (first != currentRow) {
+ //print out last row
+ out << currentRow << '\t'; //print name
+
+ //print dists
+ for (it = rowDists.begin(); it != rowDists.end(); it++) {
+ if (j >= i) { break; }
+ out << it->second << '\t';
+ j++;
+ }
+ out << endl;
+
+ //start new row
+ currentRow = first;
+ rowDists.clear();
+ rowDists[first] = 0.00;
+ rowDists[second] = dist;
+ j = 0;
+ i++;
+ }else{
+ rowDists[second] = dist;
+ }
+ }
+ //print out last row
+ out << currentRow << '\t'; //print name
+
+ //print dists
+ for (it = rowDists.begin(); it != rowDists.end(); it++) {
+ out << it->second << '\t';
+ }
+ out << endl;
+
+ in.close();
+ out.close();
+
+ remove(outfile.c_str());
+
+ return 1;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "convertToLowerTriangle");
+ exit(1);
+ }
+}
+/**************************************************************************************************/
+//its okay if the column file does not contain all the names in the fasta file, since some distance may have been above a cutoff,
+//but no sequences can be in the column file that are not in oldfasta. also, if a distance is above the cutoff given then remove it.
+//also check to make sure the 2 files have the same alignment length.
+bool DistanceCommand::sanityCheck() {
+ try{
+ bool good = true;
+
+ //make sure the 2 fasta files have the same alignment length
+ ifstream in;
+ openInputFile(fastafile, in);
+ int fastaAlignLength = 0;
+ if (in) {
+ Sequence tempIn(in);
+ fastaAlignLength = tempIn.getAligned().length();
+ }
+ in.close();
+
+ ifstream in2;
+ openInputFile(oldfastafile, in2);
+ int oldfastaAlignLength = 0;
+ if (in2) {
+ Sequence tempIn2(in2);
+ oldfastaAlignLength = tempIn2.getAligned().length();
+ }
+ in2.close();
+
+ if (fastaAlignLength != oldfastaAlignLength) { m->mothurOut("fasta files do not have the same alignment length."); m->mothurOutEndLine(); return false; }
+
+ //read fasta file and save names as well as adding them to the alignDB
+ set<string> namesOldFasta;
+
+ ifstream inFasta;
+ openInputFile(oldfastafile, inFasta);
+
+ while (!inFasta.eof()) {
+ if (m->control_pressed) { inFasta.close(); return good; }
+
+ Sequence temp(inFasta);
+
+ if (temp.getName() != "") {
+ namesOldFasta.insert(temp.getName()); //save name
+ alignDB.push_back(temp); //add to DB
+ }
+
+ gobble(inFasta);
+ }
+
+ inFasta.close();
+
+ //read through the column file checking names and removing distances above the cutoff
+ ifstream inDist;
+ openInputFile(column, inDist);
+
+ ofstream outDist;
+ string outputFile = column + ".temp";
+ openOutputFile(outputFile, outDist);
+
+ string name1, name2;
+ float dist;
+ while (!inDist.eof()) {
+ if (m->control_pressed) { inDist.close(); outDist.close(); remove(outputFile.c_str()); return good; }
+
+ inDist >> name1 >> name2 >> dist; gobble(inDist);
+
+ //both names are in fasta file and distance is below cutoff
+ if ((namesOldFasta.count(name1) == 0) || (namesOldFasta.count(name2) == 0)) { good = false; break; }
+ else{
+ if (dist <= cutoff) {
+ outDist << name1 << '\t' << name2 << '\t' << dist << endl;
+ }
+ }
+ }
+
+ inDist.close();
+ outDist.close();
+
+ if (good) {
+ remove(column.c_str());
+ rename(outputFile.c_str(), column.c_str());
+ }else{
+ remove(outputFile.c_str()); //temp file is bad because file mismatch above
+ }
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "appendFiles");