#include "onegapdist.h"
#include "onegapignore.h"
-//**********************************************************************************************************************
+//**********************************************************************************************************************
+vector<string> DistanceCommand::getValidParameters(){
+ try {
+ string Array[] = {"fasta","oldfasta","column", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir","compress"};
+ vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "getValidParameters");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+DistanceCommand::DistanceCommand(){
+ try {
+ abort = true;
+ //initialize outputTypes
+ vector<string> tempOutNames;
+ outputTypes["phylip"] = tempOutNames;
+ outputTypes["column"] = tempOutNames;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "DistanceCommand");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+vector<string> DistanceCommand::getRequiredParameters(){
+ try {
+ string Array[] = {"fasta"};
+ vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ChopSeqsCommand", "getRequiredParameters");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+vector<string> DistanceCommand::getRequiredFiles(){
+ try {
+ vector<string> myArray;
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "getRequiredFiles");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
DistanceCommand::DistanceCommand(string option) {
try {
abort = false;
else {
//valid paramters for this command
- string Array[] = {"fasta","oldfasta","column", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir"};
+ string Array[] = {"fasta","oldfasta","column", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir","compress"};
+
vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
OptionParser parser(option);
if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; }
}
+ //initialize outputTypes
+ vector<string> tempOutNames;
+ outputTypes["phylip"] = tempOutNames;
+ outputTypes["column"] = 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 = ""; }
temp = validParameter.validFile(parameters, "processors", false); if(temp == "not found"){ temp = "1"; }
convert(temp, processors);
+ temp = validParameter.validFile(parameters, "compress", false); if(temp == "not found"){ temp = "F"; }
+ convert(temp, compress);
+
output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "column"; }
if (((column != "") && (oldfastafile == "")) || ((column == "") && (oldfastafile != ""))) { m->mothurOut("If you provide column or oldfasta, you must provide both."); m->mothurOutEndLine(); abort=true; }
//**********************************************************************************************************************
-DistanceCommand::~DistanceCommand(){
-
- for(int i=0;i<lines.size();i++){
- delete lines[i];
- }
-
-}
+DistanceCommand::~DistanceCommand(){}
//**********************************************************************************************************************
void DistanceCommand::help(){
try {
m->mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n");
- m->mothurOut("The dist.seqs command parameters are fasta, oldfasta, column, calc, countends, output, cutoff and processors. \n");
+ m->mothurOut("The dist.seqs command parameters are fasta, oldfasta, column, calc, countends, output, compress, cutoff and processors. \n");
m->mothurOut("The fasta parameter is required.\n");
m->mothurOut("The oldfasta and column parameters allow you to append the distances calculated to the column file.\n");
m->mothurOut("The calc parameter allows you to specify the method of calculating the distances. Your options are: nogaps, onegap or eachgap. The default is onegap.\n");
m->mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n");
m->mothurOut("The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n");
m->mothurOut("The processors parameter allows you to specify number of processors to use. The default is 1.\n");
+ m->mothurOut("The compress parameter allows you to indicate that you want the resulting distance file compressed. The default is false.\n");
m->mothurOut("The dist.seqs command should be in the following format: \n");
m->mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n");
m->mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n");
if (output == "lt") { //does the user want lower triangle phylip formatted file
outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "phylip.dist";
- remove(outputFile.c_str());
+ remove(outputFile.c_str()); outputTypes["phylip"].push_back(outputFile);
//output numSeqs to phylip formatted dist file
}else if (output == "column") { //user wants column format
outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "dist";
+ outputTypes["column"].push_back(outputFile);
//so we don't accidentally overwrite
if (outputFile == column) {
}else { //assume square
outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "square.dist";
remove(outputFile.c_str());
+ outputTypes["phylip"].push_back(outputFile);
}
MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
//each process gets where it should start and stop in the file
- start = int (sqrt(float(pid)/float(processors)) * numSeqs);
- end = int (sqrt(float(pid+1)/float(processors)) * numSeqs);
+ if (output != "square") {
+ start = int (sqrt(float(pid)/float(processors)) * numSeqs);
+ end = int (sqrt(float(pid+1)/float(processors)) * numSeqs);
+ }else{
+ start = int ((float(pid)/float(processors)) * numSeqs);
+ end = int ((float(pid+1)/float(processors)) * numSeqs);
+ }
- if (output != "lt") {
+ if (output == "column") {
MPI_File outMPI;
int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
//do your part
string outputMyPart;
- driverMPI(start, end, outMPI, cutoff);
- if (m->control_pressed) { MPI_File_close(&outMPI); delete distCalculator; return 0; }
+ driverMPI(start, end, outMPI, cutoff);
+
+ if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
//wait on chidren
for(int i = 1; i < processors; i++) {
- if (m->control_pressed) { MPI_File_close(&outMPI); delete distCalculator; return 0; }
+ if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
- char buf[4];
- MPI_Recv(buf, 4, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
+ char buf[5];
+ MPI_Recv(buf, 5, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
}
}else { //you are a child process
//do your part
- driverMPI(start, end, outMPI, cutoff);
+ driverMPI(start, end, outMPI, cutoff);
- if (m->control_pressed) { MPI_File_close(&outMPI); delete distCalculator; return 0; }
+ if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
- char buf[4];
+ char buf[5];
strcpy(buf, "done");
//tell parent you are done.
- MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
+ MPI_Send(buf, 5, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
}
MPI_File_close(&outMPI);
//do your part
string outputMyPart;
- long mySize;
- driverMPI(start, end, outputFile, mySize);
+ unsigned long int mySize;
+
+ if (output != "square"){ driverMPI(start, end, outputFile, mySize); }
+ else { driverMPI(start, end, outputFile, mySize, output); }
- if (m->control_pressed) { delete distCalculator; return 0; }
+ if (m->control_pressed) { outputTypes.clear(); delete distCalculator; return 0; }
int amode=MPI_MODE_APPEND|MPI_MODE_WRONLY|MPI_MODE_CREATE; //
MPI_File outMPI;
//wait on chidren
for(int b = 1; b < processors; b++) {
- long fileSize;
+ unsigned long int fileSize;
- if (m->control_pressed) { MPI_File_close(&outMPI); delete distCalculator; return 0; }
+ if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
MPI_Recv(&fileSize, 1, MPI_LONG, b, tag, MPI_COMM_WORLD, &status);
MPI_File_close(&outMPI);
}else { //you are a child process
//do your part
- long size;
- driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size);
+ unsigned long int size;
+ if (output != "square"){ driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size); }
+ else { driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size, output); }
if (m->control_pressed) { delete distCalculator; return 0; }
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
//if you don't need to fork anything
if(processors == 1){
- driver(0, numSeqs, outputFile, cutoff);
+ if (output != "square") { driver(0, numSeqs, outputFile, cutoff); }
+ else { driver(0, numSeqs, outputFile, "square"); }
}else{ //you have multiple processors
- for (int i = 0; i < processors; i++) {
- lines.push_back(new linePair());
- lines[i]->start = int (sqrt(float(i)/float(processors)) * numSeqs);
- lines[i]->end = int (sqrt(float(i+1)/float(processors)) * numSeqs);
+ unsigned long int numDists = 0;
+
+ if (output == "square") {
+ numDists = numSeqs * numSeqs;
+ }else {
+ for(int i=0;i<numSeqs;i++){
+ for(int j=0;j<i;j++){
+ numDists++;
+ if (numDists > processors) { break; }
+ }
+ }
}
-
- createProcesses(outputFile);
-
- map<int, int>::iterator it = processIDS.begin();
- rename((outputFile + toString(it->second) + ".temp").c_str(), outputFile.c_str());
- it++;
- //append and remove temp files
- for (; it != processIDS.end(); it++) {
- m->appendFiles((outputFile + toString(it->second) + ".temp"), outputFile);
- remove((outputFile + toString(it->second) + ".temp").c_str());
+ if (numDists < processors) { processors = numDists; }
+
+ for (int i = 0; i < processors; i++) {
+ distlinePair tempLine;
+ lines.push_back(tempLine);
+ if (output != "square") {
+ lines[i].start = int (sqrt(float(i)/float(processors)) * numSeqs);
+ lines[i].end = int (sqrt(float(i+1)/float(processors)) * numSeqs);
+ }else{
+ lines[i].start = int ((float(i)/float(processors)) * numSeqs);
+ lines[i].end = int ((float(i+1)/float(processors)) * numSeqs);
+ }
+
}
+
+ createProcesses(outputFile);
}
#else
- ifstream inFASTA;
- driver(0, numSeqs, outputFile, cutoff);
+ //ifstream inFASTA;
+ if (output != "square") { driver(0, numSeqs, outputFile, cutoff); }
+ else { driver(0, numSeqs, outputFile, "square"); }
#endif
#endif
- if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
+ if (m->control_pressed) { outputTypes.clear(); delete distCalculator; remove(outputFile.c_str()); return 0; }
#ifdef USE_MPI
MPI_Comm_rank(MPI_COMM_WORLD, &pid);
if (pid == 0) { //only one process should output to screen
#endif
- if (output == "square") { convertMatrix(outputFile); }
+ //if (output == "square") { convertMatrix(outputFile); }
ifstream fileHandle;
fileHandle.open(outputFile.c_str());
remove(outputFile.c_str());
outputFile = column;
}
+
+ if (outputDir != "") {
+ string newOutputName = outputDir + m->getSimpleName(outputFile);
+ rename(outputFile.c_str(), newOutputName.c_str());
+ remove(outputFile.c_str());
+ outputFile = newOutputName;
+ }
}
}
#endif
- if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
+ if (m->control_pressed) { outputTypes.clear(); delete distCalculator; remove(outputFile.c_str()); return 0; }
delete distCalculator;
m->mothurOut(outputFile); m->mothurOutEndLine();
m->mothurOutEndLine();
m->mothurOut("It took " + toString(time(NULL) - startTime) + " to calculate the distances for " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
+
+
+ if (m->isTrue(compress)) {
+ m->mothurOut("Compressing..."); m->mothurOutEndLine();
+ m->mothurOut("(Replacing " + outputFile + " with " + outputFile + ".gz)"); m->mothurOutEndLine();
+ system(("gzip -v " + outputFile).c_str());
+ outputNames.push_back(outputFile + ".gz");
+ }else { outputNames.push_back(outputFile); }
+
return 0;
}
void DistanceCommand::createProcesses(string filename) {
try {
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- int process = 0;
+ int process = 1;
processIDS.clear();
//loop through and create all the processes you want
int pid = fork();
if (pid > 0) {
- processIDS[lines[process]->end] = pid; //create map from line number to pid so you can append files in correct order later
+ 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){
- driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff);
+ if (output != "square") { driver(lines[process].start, lines[process].end, filename + toString(getpid()) + ".temp", cutoff); }
+ else { driver(lines[process].start, lines[process].end, filename + toString(getpid()) + ".temp", "square"); }
+ 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);
- }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
+ }
}
-
+
+ //parent does its part
+ if (output != "square") { driver(lines[0].start, lines[0].end, filename, cutoff); }
+ else { driver(lines[0].start, lines[0].end, filename, "square"); }
+
+
//force parent to wait until all the processes are done
- for (map<int, int>::iterator it = processIDS.begin(); it != processIDS.end(); it++) {
- int temp = it->second;
+ 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((filename + toString(processIDS[i]) + ".temp"), filename);
+ remove((filename + toString(processIDS[i]) + ".temp").c_str());
+ }
#endif
}
catch(exception& e) {
if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; }
}
if (output == "lt") { outFile << dist << '\t'; }
-
- if (output == "square") { //make a square column you can convert to square phylip
- outFile << alignDB.get(i).getName() << '\t' << alignDB.get(j).getName() << '\t' << dist << endl;
- outFile << alignDB.get(j).getName() << '\t' << alignDB.get(i).getName() << '\t' << dist << endl;
- }
-
}
if (output == "lt") { outFile << endl; }
exit(1);
}
}
+/**************************************************************************************************/
+/////// need to fix to work with calcs and sequencedb
+int DistanceCommand::driver(int startLine, int endLine, string dFileName, string square){
+ try {
+
+ int startTime = time(NULL);
+
+ //column file
+ ofstream outFile(dFileName.c_str(), ios::trunc);
+ outFile.setf(ios::fixed, ios::showpoint);
+ outFile << setprecision(4);
+
+ if(startLine == 0){ outFile << alignDB.getNumSeqs() << endl; }
+
+ for(int i=startLine;i<endLine;i++){
+
+ string name = alignDB.get(i).getName();
+ //pad with spaces to make compatible
+ if (name.length() < 10) { while (name.length() < 10) { name += " "; } }
+
+ outFile << name << '\t';
+
+ for(int j=0;j<alignDB.getNumSeqs();j++){
+
+ if (m->control_pressed) { outFile.close(); return 0; }
+
+ distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
+ double dist = distCalculator->getDist();
+
+ outFile << dist << '\t';
+ }
+
+ outFile << endl;
+
+ if(i % 100 == 0){
+ m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+ }
+
+ }
+ m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
+
+ outFile.close();
+
+ return 1;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "DistanceCommand", "driver");
+ exit(1);
+ }
+}
#ifdef USE_MPI
/**************************************************************************************************/
/////// need to fix to work with calcs and sequencedb
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');
+ outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n');
}
}
}
/**************************************************************************************************/
/////// need to fix to work with calcs and sequencedb
-int DistanceCommand::driverMPI(int startLine, int endLine, string file, long& size){
+int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long int& size){
try {
MPI_Status status;
string outputString = "";
size = 0;
- if((output == "lt") && startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n"; }
+ if(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";
+
+ 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"; }
+ outputString += toString(dist) + "\t";
}
- if (output == "lt") { outputString += "\n"; }
+ outputString += "\n";
if(i % 100 == 0){
exit(1);
}
}
-#endif
/**************************************************************************************************/
+/////// need to fix to work with calcs and sequencedb
+int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long int& size, string square){
+ 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(startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n"; }
+
+ for(int i=startLine;i<endLine;i++){
+
+ 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<alignDB.getNumSeqs();j++){
+
+ if (m->control_pressed) { return 0; }
+
+ distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
+ double dist = distCalculator->getDist();
+
+ outputString += toString(dist) + "\t";
+ }
+
+ 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{
exit(1);
}
}
-/**************************************************************************************************/
+/**************************************************************************************************
int DistanceCommand::convertToLowerTriangle(string outputFile) {
try{
}
catch(exception& e) {
- m->errorOut(e, "DistanceCommand", "m->appendFiles");
- exit(1);
- }
-}
-
-/**************************************************************************************************
-void DistanceCommand::m->appendFiles(string temp, string filename) {
- try{
- ofstream output;
- ifstream input;
-
- //open output file in append mode
- m->openOutputFileAppend(filename, output);
- m->openInputFile(temp, input);
-
- while(char c = input.get()){
- if(input.eof()) { break; }
- else { output << c; }
- }
-
- input.close();
- output.close();
- }
- catch(exception& e) {
- m->errorOut(e, "DistanceCommand", "m->appendFiles");
+ m->errorOut(e, "DistanceCommand", "sanityCheck");
exit(1);
}
}
/**************************************************************************************************/
+
+