int SummaryQualCommand::createProcessesCreateSummary(vector<int>& position, vector<int>& averageQ, vector< vector<int> >& scores, string filename) {
try {
int process = 1;
int numSeqs = 0;
processIDS.clear();
bool recalc = false;
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
//loop through and create all the processes you want
while (process != processors) {
pid_t 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){
numSeqs = driverCreateSummary(position, averageQ, scores, qualfile, lines[process]);
//pass numSeqs to parent
ofstream out;
string tempFile = qualfile + m->mothurGetpid(process) + ".num.temp";
m->openOutputFile(tempFile, out);
out << numSeqs << endl;
out << position.size() << endl;
for (int k = 0; k < position.size(); k++) { out << position[k] << '\t'; } out << endl;
for (int k = 0; k < averageQ.size(); k++) { out << averageQ[k] << '\t'; } out << endl;
for (int k = 0; k < scores.size(); k++) {
for (int j = 0; j < 41; j++) {
out << scores[k][j] << '\t';
}
out << endl;
}
out << endl;
out.close();
exit(0);
}else {
m->mothurOut("[ERROR]: unable to spawn the number of processes you requested, reducing number to " + toString(process) + "\n"); processors = process;
for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
//wait to die
for (int i=0;i<processIDS.size();i++) {
int temp = processIDS[i];
wait(&temp);
}
m->control_pressed = false;
for (int i=0;i<processIDS.size();i++) {
m->mothurRemove(qualfile + (toString(processIDS[i]) + ".num.temp"));
}
recalc = true;
break;
}
}
if (recalc) {
//test line, also set recalc to true.
//for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } for (int i=0;i<processIDS.size();i++) { int temp = processIDS[i]; wait(&temp); } m->control_pressed = false; for (int i=0;i<processIDS.size();i++) {m->mothurRemove(qualfile + (toString(processIDS[i]) + ".num.temp"));}processors=3; m->mothurOut("[ERROR]: unable to spawn the number of processes you requested, reducing number to " + toString(processors) + "\n");
//redo file divide
lines.clear();
vector<unsigned long long> positions = m->divideFile(qualfile, processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(linePair(positions[i], positions[(i+1)])); }
numSeqs = 0;
processIDS.resize(0);
process = 1;
position.clear();
averageQ.clear();
scores.clear();
//loop through and create all the processes you want
while (process != processors) {
pid_t 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){
numSeqs = driverCreateSummary(position, averageQ, scores, qualfile, lines[process]);
//pass numSeqs to parent
ofstream out;
string tempFile = qualfile + m->mothurGetpid(process) + ".num.temp";
m->openOutputFile(tempFile, out);
out << numSeqs << endl;
out << position.size() << endl;
for (int k = 0; k < position.size(); k++) { out << position[k] << '\t'; } out << endl;
for (int k = 0; k < averageQ.size(); k++) { out << averageQ[k] << '\t'; } out << endl;
for (int k = 0; k < scores.size(); k++) {
for (int j = 0; j < 41; j++) {
out << scores[k][j] << '\t';
}
out << endl;
}
out << endl;
out.close();
exit(0);
}else {
m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
exit(0);
}
}
}
//do your part
numSeqs = driverCreateSummary(position, averageQ, scores, qualfile, lines[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);
}
//parent reads in and combine Filter info
for (int i = 0; i < processIDS.size(); i++) {
string tempFilename = qualfile + toString(processIDS[i]) + ".num.temp";
ifstream in;
m->openInputFile(tempFilename, in);
int temp, tempNum;
in >> tempNum; m->gobble(in); numSeqs += tempNum;
in >> tempNum; m->gobble(in);
if (position.size() < tempNum) { position.resize(tempNum, 0); }
if (averageQ.size() < tempNum) { averageQ.resize(tempNum, 0); }
if (scores.size() < tempNum) {
scores.resize(tempNum);
for (int i = 0; i < scores.size(); i++) { scores[i].resize(41, 0); }
}
for (int k = 0; k < tempNum; k++) { in >> temp; position[k] += temp; } m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; averageQ[k] += temp; } m->gobble(in);
for (int k = 0; k < tempNum; k++) {
for (int j = 0; j < 41; j++) {
in >> temp; scores[k][j] += temp;
m->gobble(in);
}
}
in.close();
m->mothurRemove(tempFilename);
}
#else
//////////////////////////////////////////////////////////////////////////////////////////////////////
//Windows version shared memory, so be careful when passing variables through the seqSumQualData struct.
//Above fork() will clone, so memory is separate, but that's not the case with windows,
//Taking advantage of shared memory to pass results vectors.
//////////////////////////////////////////////////////////////////////////////////////////////////////
vector<seqSumQualData*> pDataArray;
DWORD dwThreadIdArray[processors];
HANDLE hThreadArray[processors];
bool hasNameMap = false;
if ((namefile !="") || (countfile != "")) { hasNameMap = true; }
//Create processor worker threads.
for( int i=0; i<processors; i++ ){
// Allocate memory for thread data.
seqSumQualData* tempSum = new seqSumQualData(filename, m, lines[i].start, lines[i].end, hasNameMap, nameMap);
pDataArray.push_back(tempSum);
processIDS.push_back(i);
hThreadArray[i] = CreateThread(NULL, 0, MySeqSumQualThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]);
}
//Wait until all threads have terminated.
WaitForMultipleObjects(processors, hThreadArray, TRUE, INFINITE);
//Close all thread handles and free memory allocations.
for(int i=0; i < pDataArray.size(); i++){
numSeqs += pDataArray[i]->numSeqs;
if (pDataArray[i]->count != pDataArray[i]->end) {
m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end) + " sequences assigned to it, quitting. \n"); m->control_pressed = true;
}
int tempNum = pDataArray[i]->position.size();
if (position.size() < tempNum) { position.resize(tempNum, 0); }
if (averageQ.size() < tempNum) { averageQ.resize(tempNum, 0); }
if (scores.size() < tempNum) {
scores.resize(tempNum);
for (int i = 0; i < scores.size(); i++) { scores[i].resize(41, 0); }
}
for (int k = 0; k < tempNum; k++) { position[k] += pDataArray[i]->position[k]; }
for (int k = 0; k < tempNum; k++) { averageQ[k] += pDataArray[i]->averageQ[k]; }
for (int k = 0; k < tempNum; k++) { for (int j = 0; j < 41; j++) { scores[k][j] += pDataArray[i]->scores[k][j]; } }
CloseHandle(hThreadArray[i]);
delete pDataArray[i];
}
#endif
return numSeqs;
}
catch(exception& e) {
m->errorOut(e, "SummaryQualCommand", "createProcessesCreateSummary");
exit(1);
}
}
//***************************************************************************************************************
int SummaryQualCommand::execute(){
try{
if (abort == true) { if (calledHelp) { return 0; } return 2; }
int start = time(NULL);
int numSeqs = 0;
vector<int> position;
vector<int> averageQ;
vector< vector<int> > scores;
if (m->control_pressed) { return 0; }
if (namefile != "") { nameMap = m->readNames(namefile); }
else if (countfile != "") {
CountTable ct;
ct.readTable(countfile, false, false);
nameMap = ct.getNameMap();
}
vector<unsigned long long> positions;
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
positions = m->divideFile(qualfile, processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(linePair(positions[i], positions[(i+1)])); }
#else
if (processors == 1) {
lines.push_back(linePair(0, 1000));
}else {
positions = m->setFilePosFasta(qualfile, numSeqs);
if (numSeqs < processors) { processors = numSeqs; }
//figure out how many sequences you have to process
int numSeqsPerProcessor = numSeqs / processors;
for (int i = 0; i < processors; i++) {
int startIndex = i * numSeqsPerProcessor;
if(i == (processors - 1)){ numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor; }
lines.push_back(linePair(positions[startIndex], numSeqsPerProcessor));
}
}
#endif
if(processors == 1){ numSeqs = driverCreateSummary(position, averageQ, scores, qualfile, lines[0]); }
else{ numSeqs = createProcessesCreateSummary(position, averageQ, scores, qualfile); }
if (m->control_pressed) { return 0; }
//print summary file
map<string, string> variables;
variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(qualfile));
string summaryFile = getOutputFileName("summary",variables);
printQual(summaryFile, position, averageQ, scores);
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
//output results to screen
cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint);
m->mothurOutEndLine();
m->mothurOut("Position\tNumSeqs\tAverageQ"); m->mothurOutEndLine();
for (int i = 0; i < position.size(); i+=100) {
float average = averageQ[i] / (float) position[i];
cout << i << '\t' << position[i] << '\t' << average;
m->mothurOutJustToLog(toString(i) + "\t" + toString(position[i]) + "\t" + toString(average)); m->mothurOutEndLine();
}
m->mothurOutEndLine();
m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
m->mothurOutEndLine();
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
m->mothurOut(summaryFile); m->mothurOutEndLine(); outputNames.push_back(summaryFile); outputTypes["summary"].push_back(summaryFile);
m->mothurOutEndLine();
return 0;
}
catch(exception& e) {
m->errorOut(e, "SummaryQualCommand", "execute");
exit(1);
}
}
int SeqSummaryCommand::createProcessesCreateSummary(vector<int>& startPosition, vector<int>& endPosition, vector<int>& seqLength, vector<int>& ambigBases, vector<int>& longHomoPolymer, string filename, string sumFile) {
try {
int process = 1;
int num = 0;
processIDS.clear();
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
//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){
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, sumFile + toString(getpid()) + ".temp", lines[process]);
//pass numSeqs to parent
ofstream out;
string tempFile = fastafile + toString(getpid()) + ".num.temp";
m->openOutputFile(tempFile, out);
out << num << endl;
out << startPosition.size() << endl;
for (int k = 0; k < startPosition.size(); k++) { out << startPosition[k] << '\t'; } out << endl;
for (int k = 0; k < endPosition.size(); k++) { out << endPosition[k] << '\t'; } out << endl;
for (int k = 0; k < seqLength.size(); k++) { out << seqLength[k] << '\t'; } out << endl;
for (int k = 0; k < ambigBases.size(); k++) { out << ambigBases[k] << '\t'; } out << endl;
for (int k = 0; k < longHomoPolymer.size(); k++) { out << longHomoPolymer[k] << '\t'; } out << endl;
out.close();
exit(0);
}else {
m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
exit(0);
}
}
//do your part
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, sumFile, lines[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);
}
//parent reads in and combine Filter info
for (int i = 0; i < processIDS.size(); i++) {
string tempFilename = fastafile + toString(processIDS[i]) + ".num.temp";
ifstream in;
m->openInputFile(tempFilename, in);
int temp, tempNum;
in >> tempNum; m->gobble(in); num += tempNum;
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; startPosition.push_back(temp); } m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; endPosition.push_back(temp); } m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; seqLength.push_back(temp); } m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; ambigBases.push_back(temp); } m->gobble(in);
for (int k = 0; k < tempNum; k++) { in >> temp; longHomoPolymer.push_back(temp); } m->gobble(in);
in.close();
m->mothurRemove(tempFilename);
m->appendFiles((sumFile + toString(processIDS[i]) + ".temp"), sumFile);
m->mothurRemove((sumFile + toString(processIDS[i]) + ".temp"));
}
#else
//////////////////////////////////////////////////////////////////////////////////////////////////////
//Windows version shared memory, so be careful when passing variables through the seqSumData struct.
//Above fork() will clone, so memory is separate, but that's not the case with windows,
//Taking advantage of shared memory to allow both threads to add info to vectors.
//////////////////////////////////////////////////////////////////////////////////////////////////////
vector<seqSumData*> pDataArray;
DWORD dwThreadIdArray[processors-1];
HANDLE hThreadArray[processors-1];
bool hasNameMap = false;
if ((namefile !="") || (countfile != "")) { hasNameMap = true; }
//Create processor worker threads.
for( int i=0; i<processors-1; i++ ){
string extension = "";
if (i != 0) { extension = toString(i) + ".temp"; processIDS.push_back(i); }
// Allocate memory for thread data.
seqSumData* tempSum = new seqSumData(filename, (sumFile+extension), m, lines[i]->start, lines[i]->end, hasNameMap, nameMap);
pDataArray.push_back(tempSum);
//MySeqSumThreadFunction is in header. It must be global or static to work with the threads.
//default security attributes, thread function name, argument to thread function, use default creation flags, returns the thread identifier
hThreadArray[i] = CreateThread(NULL, 0, MySeqSumThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]);
}
//do your part
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, (sumFile+toString(processors-1)+".temp"), lines[processors-1]);
processIDS.push_back(processors-1);
//Wait until all threads have terminated.
WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
//Close all thread handles and free memory allocations.
for(int i=0; i < pDataArray.size(); i++){
num += pDataArray[i]->count;
if (pDataArray[i]->count != pDataArray[i]->end) {
m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end) + " sequences assigned to it, quitting. \n"); m->control_pressed = true;
}
for (int k = 0; k < pDataArray[i]->startPosition.size(); k++) { startPosition.push_back(pDataArray[i]->startPosition[k]); }
for (int k = 0; k < pDataArray[i]->endPosition.size(); k++) { endPosition.push_back(pDataArray[i]->endPosition[k]); }
for (int k = 0; k < pDataArray[i]->seqLength.size(); k++) { seqLength.push_back(pDataArray[i]->seqLength[k]); }
for (int k = 0; k < pDataArray[i]->ambigBases.size(); k++) { ambigBases.push_back(pDataArray[i]->ambigBases[k]); }
for (int k = 0; k < pDataArray[i]->longHomoPolymer.size(); k++) { longHomoPolymer.push_back(pDataArray[i]->longHomoPolymer[k]); }
CloseHandle(hThreadArray[i]);
delete pDataArray[i];
}
//append files
for(int i=0;i<processIDS.size();i++){
m->appendFiles((sumFile + toString(processIDS[i]) + ".temp"), sumFile);
m->mothurRemove((sumFile + toString(processIDS[i]) + ".temp"));
}
#endif
return num;
}
catch(exception& e) {
m->errorOut(e, "SeqSummaryCommand", "createProcessesCreateSummary");
exit(1);
}
}
int SeqSummaryCommand::execute(){
try{
if (abort == true) { if (calledHelp) { return 0; } return 2; }
//set current fasta to fastafile
m->setFastaFile(fastafile);
map<string, string> variables;
variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(fastafile));
string summaryFile = getOutputFileName("summary",variables);
int numSeqs = 0;
vector<int> startPosition;
vector<int> endPosition;
vector<int> seqLength;
vector<int> ambigBases;
vector<int> longHomoPolymer;
if (namefile != "") { nameMap = m->readNames(namefile); }
else if (countfile != "") {
CountTable ct;
ct.readTable(countfile, false, false);
nameMap = ct.getNameMap();
}
if (m->control_pressed) { return 0; }
#ifdef USE_MPI
int pid, numSeqsPerProcessor;
int tag = 2001;
int startTag = 1; int endTag = 2; int lengthTag = 3; int baseTag = 4; int lhomoTag = 5;
int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
vector<unsigned long long> MPIPos;
MPI_Status status;
MPI_Status statusOut;
MPI_File inMPI;
MPI_File outMPI;
MPI_Comm_size(MPI_COMM_WORLD, &processors);
MPI_Comm_rank(MPI_COMM_WORLD, &pid);
char tempFileName[1024];
strcpy(tempFileName, fastafile.c_str());
char sumFileName[1024];
strcpy(sumFileName, summaryFile.c_str());
MPI_File_open(MPI_COMM_WORLD, tempFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
MPI_File_open(MPI_COMM_WORLD, sumFileName, outMode, MPI_INFO_NULL, &outMPI);
if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); return 0; }
if (pid == 0) { //you are the root process
//print header
string outputString = "seqname\tstart\tend\tnbases\tambigs\tpolymer\tnumSeqs\n";
int length = outputString.length();
char* buf2 = new char[length];
memcpy(buf2, outputString.c_str(), length);
MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &statusOut);
delete buf2;
MPIPos = m->setFilePosFasta(fastafile, numSeqs); //fills MPIPos, returns numSeqs
for(int i = 1; i < processors; i++) {
MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
MPI_Send(&MPIPos[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
}
//figure out how many sequences you have to do
numSeqsPerProcessor = numSeqs / processors;
int startIndex = pid * numSeqsPerProcessor;
if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
//do your part
MPICreateSummary(startIndex, numSeqsPerProcessor, startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, inMPI, outMPI, MPIPos);
}else { //i am the child process
MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
MPIPos.resize(numSeqs+1);
MPI_Recv(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
//figure out how many sequences you have to align
numSeqsPerProcessor = numSeqs / processors;
int startIndex = pid * numSeqsPerProcessor;
if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
//do your part
MPICreateSummary(startIndex, numSeqsPerProcessor, startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, inMPI, outMPI, MPIPos);
}
MPI_File_close(&inMPI);
MPI_File_close(&outMPI);
MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
if (pid == 0) {
//get the info from the child processes
for(int i = 1; i < processors; i++) {
int size;
MPI_Recv(&size, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
vector<int> temp; temp.resize(size+1);
for(int j = 0; j < 5; j++) {
MPI_Recv(&temp[0], (size+1), MPI_INT, i, 2001, MPI_COMM_WORLD, &status);
int receiveTag = temp[temp.size()-1]; //child process added a int to the end to indicate what count this is for
if (receiveTag == startTag) {
for (int k = 0; k < size; k++) { startPosition.push_back(temp[k]); }
}else if (receiveTag == endTag) {
for (int k = 0; k < size; k++) { endPosition.push_back(temp[k]); }
}else if (receiveTag == lengthTag) {
for (int k = 0; k < size; k++) { seqLength.push_back(temp[k]); }
}else if (receiveTag == baseTag) {
for (int k = 0; k < size; k++) { ambigBases.push_back(temp[k]); }
}else if (receiveTag == lhomoTag) {
for (int k = 0; k < size; k++) { longHomoPolymer.push_back(temp[k]); }
}
}
}
}else{
//send my counts
int size = startPosition.size();
MPI_Send(&size, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
startPosition.push_back(startTag);
int ierr = MPI_Send(&(startPosition[0]), (size+1), MPI_INT, 0, 2001, MPI_COMM_WORLD);
endPosition.push_back(endTag);
ierr = MPI_Send (&(endPosition[0]), (size+1), MPI_INT, 0, 2001, MPI_COMM_WORLD);
seqLength.push_back(lengthTag);
ierr = MPI_Send(&(seqLength[0]), (size+1), MPI_INT, 0, 2001, MPI_COMM_WORLD);
ambigBases.push_back(baseTag);
ierr = MPI_Send(&(ambigBases[0]), (size+1), MPI_INT, 0, 2001, MPI_COMM_WORLD);
longHomoPolymer.push_back(lhomoTag);
ierr = MPI_Send(&(longHomoPolymer[0]), (size+1), MPI_INT, 0, 2001, MPI_COMM_WORLD);
}
MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
#else
vector<unsigned long long> positions;
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
positions = m->divideFile(fastafile, processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); }
#else
positions = m->setFilePosFasta(fastafile, numSeqs);
if (positions.size() < processors) { processors = positions.size(); }
//figure out how many sequences you have to process
int numSeqsPerProcessor = numSeqs / processors;
for (int i = 0; i < processors; i++) {
int startIndex = i * numSeqsPerProcessor;
if(i == (processors - 1)){ numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor; }
lines.push_back(new linePair(positions[startIndex], numSeqsPerProcessor));
}
#endif
if(processors == 1){
numSeqs = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, summaryFile, lines[0]);
}else{
numSeqs = createProcessesCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, summaryFile);
}
if (m->control_pressed) { return 0; }
#endif
#ifdef USE_MPI
if (pid == 0) {
#endif
sort(startPosition.begin(), startPosition.end());
sort(endPosition.begin(), endPosition.end());
sort(seqLength.begin(), seqLength.end());
sort(ambigBases.begin(), ambigBases.end());
sort(longHomoPolymer.begin(), longHomoPolymer.end());
int size = startPosition.size();
//find means
unsigned long long meanStartPosition, meanEndPosition, meanSeqLength, meanAmbigBases, meanLongHomoPolymer;
meanStartPosition = 0; meanEndPosition = 0; meanSeqLength = 0; meanAmbigBases = 0; meanLongHomoPolymer = 0;
for (int i = 0; i < size; i++) {
meanStartPosition += startPosition[i];
meanEndPosition += endPosition[i];
meanSeqLength += seqLength[i];
meanAmbigBases += ambigBases[i];
meanLongHomoPolymer += longHomoPolymer[i];
}
double meanstartPosition, meanendPosition, meanseqLength, meanambigBases, meanlongHomoPolymer;
meanstartPosition = meanStartPosition / (double) size; meanendPosition = meanEndPosition /(double) size; meanlongHomoPolymer = meanLongHomoPolymer / (double) size; meanseqLength = meanSeqLength / (double) size; meanambigBases = meanAmbigBases /(double) size;
int ptile0_25 = int(size * 0.025);
int ptile25 = int(size * 0.250);
int ptile50 = int(size * 0.500);
int ptile75 = int(size * 0.750);
int ptile97_5 = int(size * 0.975);
int ptile100 = size - 1;
//to compensate for blank sequences that would result in startPosition and endPostion equalling -1
if (startPosition[0] == -1) { startPosition[0] = 0; }
if (endPosition[0] == -1) { endPosition[0] = 0; }
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
m->mothurOutEndLine();
m->mothurOut("\t\tStart\tEnd\tNBases\tAmbigs\tPolymer\tNumSeqs"); m->mothurOutEndLine();
m->mothurOut("Minimum:\t" + toString(startPosition[0]) + "\t" + toString(endPosition[0]) + "\t" + toString(seqLength[0]) + "\t" + toString(ambigBases[0]) + "\t" + toString(longHomoPolymer[0]) + "\t" + toString(1)); m->mothurOutEndLine();
m->mothurOut("2.5%-tile:\t" + toString(startPosition[ptile0_25]) + "\t" + toString(endPosition[ptile0_25]) + "\t" + toString(seqLength[ptile0_25]) + "\t" + toString(ambigBases[ptile0_25]) + "\t"+ toString(longHomoPolymer[ptile0_25]) + "\t" + toString(ptile0_25+1)); m->mothurOutEndLine();
m->mothurOut("25%-tile:\t" + toString(startPosition[ptile25]) + "\t" + toString(endPosition[ptile25]) + "\t" + toString(seqLength[ptile25]) + "\t" + toString(ambigBases[ptile25]) + "\t" + toString(longHomoPolymer[ptile25]) + "\t" + toString(ptile25+1)); m->mothurOutEndLine();
m->mothurOut("Median: \t" + toString(startPosition[ptile50]) + "\t" + toString(endPosition[ptile50]) + "\t" + toString(seqLength[ptile50]) + "\t" + toString(ambigBases[ptile50]) + "\t" + toString(longHomoPolymer[ptile50]) + "\t" + toString(ptile50+1)); m->mothurOutEndLine();
m->mothurOut("75%-tile:\t" + toString(startPosition[ptile75]) + "\t" + toString(endPosition[ptile75]) + "\t" + toString(seqLength[ptile75]) + "\t" + toString(ambigBases[ptile75]) + "\t" + toString(longHomoPolymer[ptile75]) + "\t" + toString(ptile75+1)); m->mothurOutEndLine();
m->mothurOut("97.5%-tile:\t" + toString(startPosition[ptile97_5]) + "\t" + toString(endPosition[ptile97_5]) + "\t" + toString(seqLength[ptile97_5]) + "\t" + toString(ambigBases[ptile97_5]) + "\t" + toString(longHomoPolymer[ptile97_5]) + "\t" + toString(ptile97_5+1)); m->mothurOutEndLine();
m->mothurOut("Maximum:\t" + toString(startPosition[ptile100]) + "\t" + toString(endPosition[ptile100]) + "\t" + toString(seqLength[ptile100]) + "\t" + toString(ambigBases[ptile100]) + "\t" + toString(longHomoPolymer[ptile100]) + "\t" + toString(ptile100+1)); m->mothurOutEndLine();
m->mothurOut("Mean:\t" + toString(meanstartPosition) + "\t" + toString(meanendPosition) + "\t" + toString(meanseqLength) + "\t" + toString(meanambigBases) + "\t" + toString(meanlongHomoPolymer)); m->mothurOutEndLine();
if ((namefile == "") && (countfile == "")) { m->mothurOut("# of Seqs:\t" + toString(numSeqs)); m->mothurOutEndLine(); }
else { m->mothurOut("# of unique seqs:\t" + toString(numSeqs)); m->mothurOutEndLine(); m->mothurOut("total # of seqs:\t" + toString(startPosition.size())); m->mothurOutEndLine(); }
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
m->mothurOutEndLine();
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
m->mothurOut(summaryFile); m->mothurOutEndLine(); outputNames.push_back(summaryFile); outputTypes["summary"].push_back(summaryFile);
m->mothurOutEndLine();
#ifdef USE_MPI
}
#endif
//set fasta file as new current fastafile
string current = "";
itTypes = outputTypes.find("summary");
if (itTypes != outputTypes.end()) {
if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setSummaryFile(current); }
}
return 0;
}
catch(exception& e) {
m->errorOut(e, "SeqSummaryCommand", "execute");
exit(1);
}
}
long long SeqSummaryCommand::createProcessesCreateSummary(map<int, long long>& startPosition, map<int, long long>& endPosition, map<int, long long>& seqLength, map<int, long long>& ambigBases, map<int, long long>& longHomoPolymer, string filename, string sumFile) {
try {
int process = 1;
int num = 0;
processIDS.clear();
bool recalc = false;
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
//loop through and create all the processes you want
while (process != processors) {
pid_t 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){
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, sumFile + m->mothurGetpid(process) + ".temp", lines[process]);
//pass numSeqs to parent
ofstream out;
string tempFile = fastafile + m->mothurGetpid(process) + ".num.temp";
m->openOutputFile(tempFile, out);
out << num << endl;
out << startPosition.size() << endl;
for (map<int, long long>::iterator it = startPosition.begin(); it != startPosition.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << endPosition.size() << endl;
for (map<int, long long>::iterator it = endPosition.begin(); it != endPosition.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << seqLength.size() << endl;
for (map<int, long long>::iterator it = seqLength.begin(); it != seqLength.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << ambigBases.size() << endl;
for (map<int, long long>::iterator it = ambigBases.begin(); it != ambigBases.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << longHomoPolymer.size() << endl;
for (map<int, long long>::iterator it = longHomoPolymer.begin(); it != longHomoPolymer.end(); it++) { out << it->first << '\t' << it->second << endl; }
out.close();
exit(0);
}else {
m->mothurOut("[ERROR]: unable to spawn the number of processes you requested, reducing number to " + toString(process) + "\n"); processors = process;
for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
//wait to die
for (int i=0;i<processIDS.size();i++) {
int temp = processIDS[i];
wait(&temp);
}
m->control_pressed = false;
for (int i=0;i<processIDS.size();i++) {
m->mothurRemove(fastafile + (toString(processIDS[i]) + ".num.temp"));
m->mothurRemove(sumFile + (toString(processIDS[i]) + ".temp"));
}
recalc = true;
break;
}
}
if (recalc) {
//test line, also set recalc to true.
//for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } for (int i=0;i<processIDS.size();i++) { int temp = processIDS[i]; wait(&temp); } m->control_pressed = false; for (int i=0;i<processIDS.size();i++) {m->mothurRemove(fastafile + (toString(processIDS[i]) + ".num.temp"));}processors=3; m->mothurOut("[ERROR]: unable to spawn the number of processes you requested, reducing number to " + toString(processors) + "\n");
//redo file divide
for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
vector<unsigned long long> positions = m->divideFile(filename, processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); }
startPosition.clear();
endPosition.clear();
seqLength.clear();
ambigBases.clear();
longHomoPolymer.clear();
num = 0;
processIDS.resize(0);
process = 1;
//loop through and create all the processes you want
while (process != processors) {
pid_t 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){
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, sumFile + m->mothurGetpid(process) + ".temp", lines[process]);
//pass numSeqs to parent
ofstream out;
string tempFile = fastafile + m->mothurGetpid(process) + ".num.temp";
m->openOutputFile(tempFile, out);
out << num << endl;
out << startPosition.size() << endl;
for (map<int, long long>::iterator it = startPosition.begin(); it != startPosition.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << endPosition.size() << endl;
for (map<int, long long>::iterator it = endPosition.begin(); it != endPosition.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << seqLength.size() << endl;
for (map<int, long long>::iterator it = seqLength.begin(); it != seqLength.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << ambigBases.size() << endl;
for (map<int, long long>::iterator it = ambigBases.begin(); it != ambigBases.end(); it++) { out << it->first << '\t' << it->second << endl; }
out << longHomoPolymer.size() << endl;
for (map<int, long long>::iterator it = longHomoPolymer.begin(); it != longHomoPolymer.end(); it++) { out << it->first << '\t' << it->second << endl; }
out.close();
exit(0);
}else {
m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
exit(0);
}
}
}
//do your part
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, sumFile, lines[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);
}
//parent reads in and combine Filter info
for (int i = 0; i < processIDS.size(); i++) {
string tempFilename = fastafile + toString(processIDS[i]) + ".num.temp";
ifstream in;
m->openInputFile(tempFilename, in);
long long tempNum;
in >> tempNum; m->gobble(in); num += tempNum;
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) {
long long first, second;
in >> first; m->gobble(in); in >> second; m->gobble(in);
map<int, long long>::iterator it = startPosition.find(first);
if (it == startPosition.end()) { startPosition[first] = second; } //first finding of this start position, set count.
else { it->second += second; } //add counts
}
m->gobble(in);
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) {
long long first, second;
in >> first; m->gobble(in); in >> second; m->gobble(in);
map<int, long long>::iterator it = endPosition.find(first);
if (it == endPosition.end()) { endPosition[first] = second; } //first finding of this end position, set count.
else { it->second += second; } //add counts
}
m->gobble(in);
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) {
long long first, second;
in >> first; m->gobble(in); in >> second; m->gobble(in);
map<int, long long>::iterator it = seqLength.find(first);
if (it == seqLength.end()) { seqLength[first] = second; } //first finding of this end position, set count.
else { it->second += second; } //add counts
}
m->gobble(in);
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) {
long long first, second;
in >> first; m->gobble(in); in >> second; m->gobble(in);
map<int, long long>::iterator it = ambigBases.find(first);
if (it == ambigBases.end()) { ambigBases[first] = second; } //first finding of this end position, set count.
else { it->second += second; } //add counts
}
m->gobble(in);
in >> tempNum; m->gobble(in);
for (int k = 0; k < tempNum; k++) {
long long first, second;
in >> first; m->gobble(in); in >> second; m->gobble(in);
map<int, long long>::iterator it = longHomoPolymer.find(first);
if (it == longHomoPolymer.end()) { longHomoPolymer[first] = second; } //first finding of this end position, set count.
else { it->second += second; } //add counts
}
m->gobble(in);
in.close();
m->mothurRemove(tempFilename);
m->appendFiles((sumFile + toString(processIDS[i]) + ".temp"), sumFile);
m->mothurRemove((sumFile + toString(processIDS[i]) + ".temp"));
}
#else
//////////////////////////////////////////////////////////////////////////////////////////////////////
//Windows version shared memory, so be careful when passing variables through the seqSumData struct.
//Above fork() will clone, so memory is separate, but that's not the case with windows,
//Taking advantage of shared memory to allow both threads to add info to vectors.
//////////////////////////////////////////////////////////////////////////////////////////////////////
vector<seqSumData*> pDataArray;
DWORD dwThreadIdArray[processors-1];
HANDLE hThreadArray[processors-1];
bool hasNameMap = false;
if ((namefile !="") || (countfile != "")) { hasNameMap = true; }
//Create processor worker threads.
for( int i=0; i<processors-1; i++ ){
string extension = "";
if (i != 0) { extension = toString(i) + ".temp"; processIDS.push_back(i); }
// Allocate memory for thread data.
seqSumData* tempSum = new seqSumData(filename, (sumFile+extension), m, lines[i]->start, lines[i]->end, hasNameMap, nameMap);
pDataArray.push_back(tempSum);
//MySeqSumThreadFunction is in header. It must be global or static to work with the threads.
//default security attributes, thread function name, argument to thread function, use default creation flags, returns the thread identifier
hThreadArray[i] = CreateThread(NULL, 0, MySeqSumThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]);
}
//do your part
num = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, (sumFile+toString(processors-1)+".temp"), lines[processors-1]);
processIDS.push_back(processors-1);
//Wait until all threads have terminated.
WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
//Close all thread handles and free memory allocations.
for(int i=0; i < pDataArray.size(); i++){
num += pDataArray[i]->count;
if (pDataArray[i]->count != pDataArray[i]->end) {
m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end) + " sequences assigned to it, quitting. \n"); m->control_pressed = true;
}
for (map<int, long long>::iterator it = pDataArray[i]->startPosition.begin(); it != pDataArray[i]->startPosition.end(); it++) {
map<int, long long>::iterator itMain = startPosition.find(it->first);
if (itMain == startPosition.end()) { //newValue
startPosition[it->first] = it->second;
}else { itMain->second += it->second; } //merge counts
}
for (map<int, long long>::iterator it = pDataArray[i]->endPosition.begin(); it != pDataArray[i]->endPosition.end(); it++) {
map<int, long long>::iterator itMain = endPosition.find(it->first);
if (itMain == endPosition.end()) { //newValue
endPosition[it->first] = it->second;
}else { itMain->second += it->second; } //merge counts
}
for (map<int, long long>::iterator it = pDataArray[i]->seqLength.begin(); it != pDataArray[i]->seqLength.end(); it++) {
map<int, long long>::iterator itMain = seqLength.find(it->first);
if (itMain == seqLength.end()) { //newValue
seqLength[it->first] = it->second;
}else { itMain->second += it->second; } //merge counts
}
for (map<int, long long>::iterator it = pDataArray[i]->ambigBases.begin(); it != pDataArray[i]->ambigBases.end(); it++) {
map<int, long long>::iterator itMain = ambigBases.find(it->first);
if (itMain == ambigBases.end()) { //newValue
ambigBases[it->first] = it->second;
}else { itMain->second += it->second; } //merge counts
}
for (map<int, long long>::iterator it = pDataArray[i]->longHomoPolymer.begin(); it != pDataArray[i]->longHomoPolymer.end(); it++) {
map<int, long long>::iterator itMain = longHomoPolymer.find(it->first);
if (itMain == longHomoPolymer.end()) { //newValue
longHomoPolymer[it->first] = it->second;
}else { itMain->second += it->second; } //merge counts
}
CloseHandle(hThreadArray[i]);
delete pDataArray[i];
}
//append files
for(int i=0;i<processIDS.size();i++){
m->appendFiles((sumFile + toString(processIDS[i]) + ".temp"), sumFile);
m->mothurRemove((sumFile + toString(processIDS[i]) + ".temp"));
}
#endif
return num;
}
catch(exception& e) {
m->errorOut(e, "SeqSummaryCommand", "createProcessesCreateSummary");
exit(1);
}
}
int SeqSummaryCommand::execute(){
try{
if (abort == true) { if (calledHelp) { return 0; } return 2; }
int start = time(NULL);
//set current fasta to fastafile
m->setFastaFile(fastafile);
map<string, string> variables;
variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(fastafile));
string summaryFile = getOutputFileName("summary",variables);
long long numSeqs = 0;
long long size = 0;
long long numUniques = 0;
map<int, long long> startPosition;
map<int, long long> endPosition;
map<int, long long> seqLength;
map<int, long long> ambigBases;
map<int, long long> longHomoPolymer;
if (namefile != "") { nameMap = m->readNames(namefile); numUniques = nameMap.size(); }
else if (countfile != "") {
CountTable ct;
ct.readTable(countfile, false, false);
nameMap = ct.getNameMap();
size = ct.getNumSeqs();
numUniques = ct.getNumUniqueSeqs();
}
if (m->control_pressed) { return 0; }
vector<unsigned long long> positions;
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
positions = m->divideFile(fastafile, processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); }
#else
positions = m->setFilePosFasta(fastafile, numSeqs);
if (numSeqs < processors) { processors = numSeqs; }
//figure out how many sequences you have to process
int numSeqsPerProcessor = numSeqs / processors;
for (int i = 0; i < processors; i++) {
int startIndex = i * numSeqsPerProcessor;
if(i == (processors - 1)){ numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor; }
lines.push_back(new linePair(positions[startIndex], numSeqsPerProcessor));
}
#endif
if(processors == 1){
numSeqs = driverCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, summaryFile, lines[0]);
}else{
numSeqs = createProcessesCreateSummary(startPosition, endPosition, seqLength, ambigBases, longHomoPolymer, fastafile, summaryFile);
}
if (m->control_pressed) { return 0; }
//set size
if (countfile != "") {}//already set
else if (namefile == "") { size = numSeqs; }
else { for (map<int, long long>::iterator it = startPosition.begin(); it != startPosition.end(); it++) { size += it->second; } }
if ((namefile != "") || (countfile != "")) {
string type = "count";
if (namefile != "") { type = "name"; }
if (numSeqs != numUniques) { // do fasta and name/count files match
m->mothurOut("[ERROR]: Your " + type + " file contains " + toString(numUniques) + " unique sequences, but your fasta file contains " + toString(numSeqs) + ". File mismatch detected, quitting command.\n"); m->control_pressed = true;
}
}
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
long long ptile0_25 = 1+(long long)(size * 0.025); //number of sequences at 2.5%
long long ptile25 = 1+(long long)(size * 0.250); //number of sequences at 25%
long long ptile50 = 1+(long long)(size * 0.500);
long long ptile75 = 1+(long long)(size * 0.750);
long long ptile97_5 = 1+(long long)(size * 0.975);
long long ptile100 = (long long)(size);
vector<int> starts; starts.resize(7,0); vector<int> ends; ends.resize(7,0); vector<int> ambigs; ambigs.resize(7,0); vector<int> lengths; lengths.resize(7,0); vector<int> homops; homops.resize(7,0);
//find means
long long meanStartPosition, meanEndPosition, meanSeqLength, meanAmbigBases, meanLongHomoPolymer;
meanStartPosition = 0; meanEndPosition = 0; meanSeqLength = 0; meanAmbigBases = 0; meanLongHomoPolymer = 0;
//minimum
if ((startPosition.begin())->first == -1) { starts[0] = 0; }
else {starts[0] = (startPosition.begin())->first; }
long long totalSoFar = 0;
//set all values to min
starts[1] = starts[0]; starts[2] = starts[0]; starts[3] = starts[0]; starts[4] = starts[0]; starts[5] = starts[0];
int lastValue = 0;
for (map<int, long long>::iterator it = startPosition.begin(); it != startPosition.end(); it++) {
int value = it->first; if (value == -1) { value = 0; }
meanStartPosition += (value*it->second);
totalSoFar += it->second;
if (((totalSoFar <= ptile0_25) && (totalSoFar > 1)) || ((lastValue < ptile0_25) && (totalSoFar > ptile0_25))){ starts[1] = value; } //save value
if (((totalSoFar <= ptile25) && (totalSoFar > ptile0_25)) || ((lastValue < ptile25) && (totalSoFar > ptile25))) { starts[2] = value; } //save value
if (((totalSoFar <= ptile50) && (totalSoFar > ptile25)) || ((lastValue < ptile50) && (totalSoFar > ptile50))) { starts[3] = value; } //save value
if (((totalSoFar <= ptile75) && (totalSoFar > ptile50)) || ((lastValue < ptile75) && (totalSoFar > ptile75))) { starts[4] = value; } //save value
if (((totalSoFar <= ptile97_5) && (totalSoFar > ptile75)) || ((lastValue < ptile97_5) && (totalSoFar > ptile97_5))) { starts[5] = value; } //save value
if ((totalSoFar <= ptile100) && (totalSoFar > ptile97_5)) { starts[6] = value; } //save value
lastValue = totalSoFar;
}
starts[6] = (startPosition.rbegin())->first;
if ((endPosition.begin())->first == -1) { ends[0] = 0; }
else {ends[0] = (endPosition.begin())->first; }
totalSoFar = 0;
//set all values to min
ends[1] = ends[0]; ends[2] = ends[0]; ends[3] = ends[0]; ends[4] = ends[0]; ends[5] = ends[0];
lastValue = 0;
for (map<int, long long>::iterator it = endPosition.begin(); it != endPosition.end(); it++) {
int value = it->first; if (value == -1) { value = 0; }
meanEndPosition += (value*it->second);
totalSoFar += it->second;
if (((totalSoFar <= ptile0_25) && (totalSoFar > 1)) || ((lastValue < ptile0_25) && (totalSoFar > ptile0_25))){ ends[1] = value; } //save value
if (((totalSoFar <= ptile25) && (totalSoFar > ptile0_25)) || ((lastValue < ptile25) && (totalSoFar > ptile25))) { ends[2] = value; } //save value
if (((totalSoFar <= ptile50) && (totalSoFar > ptile25)) || ((lastValue < ptile50) && (totalSoFar > ptile50))) { ends[3] = value; } //save value
if (((totalSoFar <= ptile75) && (totalSoFar > ptile50)) || ((lastValue < ptile75) && (totalSoFar > ptile75))) { ends[4] = value; } //save value
if (((totalSoFar <= ptile97_5) && (totalSoFar > ptile75)) || ((lastValue < ptile97_5) && (totalSoFar > ptile97_5))) { ends[5] = value; } //save value
if ((totalSoFar <= ptile100) && (totalSoFar > ptile97_5)) { ends[6] = value; } //save value
lastValue = totalSoFar;
}
ends[6] = (endPosition.rbegin())->first;
if ((seqLength.begin())->first == -1) { lengths[0] = 0; }
else {lengths[0] = (seqLength.begin())->first; }
//set all values to min
lengths[1] = lengths[0]; lengths[2] = lengths[0]; lengths[3] = lengths[0]; lengths[4] = lengths[0]; lengths[5] = lengths[0];
totalSoFar = 0;
lastValue = 0;
for (map<int, long long>::iterator it = seqLength.begin(); it != seqLength.end(); it++) {
int value = it->first;
meanSeqLength += (value*it->second);
totalSoFar += it->second;
if (((totalSoFar <= ptile0_25) && (totalSoFar > 1)) || ((lastValue < ptile0_25) && (totalSoFar > ptile0_25))){ lengths[1] = value; } //save value
if (((totalSoFar <= ptile25) && (totalSoFar > ptile0_25)) || ((lastValue < ptile25) && (totalSoFar > ptile25))) { lengths[2] = value; } //save value
if (((totalSoFar <= ptile50) && (totalSoFar > ptile25)) || ((lastValue < ptile50) && (totalSoFar > ptile50))) { lengths[3] = value; } //save value
if (((totalSoFar <= ptile75) && (totalSoFar > ptile50)) || ((lastValue < ptile75) && (totalSoFar > ptile75))) { lengths[4] = value; } //save value
if (((totalSoFar <= ptile97_5) && (totalSoFar > ptile75)) || ((lastValue < ptile97_5) && (totalSoFar > ptile97_5))) { lengths[5] = value; } //save value
if ((totalSoFar <= ptile100) && (totalSoFar > ptile97_5)) { lengths[6] = value; } //save value
lastValue = totalSoFar;
}
lengths[6] = (seqLength.rbegin())->first;
if ((ambigBases.begin())->first == -1) { ambigs[0] = 0; }
else {ambigs[0] = (ambigBases.begin())->first; }
//set all values to min
ambigs[1] = ambigs[0]; ambigs[2] = ambigs[0]; ambigs[3] = ambigs[0]; ambigs[4] = ambigs[0]; ambigs[5] = ambigs[0];
totalSoFar = 0;
lastValue = 0;
for (map<int, long long>::iterator it = ambigBases.begin(); it != ambigBases.end(); it++) {
int value = it->first;
meanAmbigBases += (value*it->second);
totalSoFar += it->second;
if (((totalSoFar <= ptile0_25) && (totalSoFar > 1)) || ((lastValue < ptile0_25) && (totalSoFar > ptile0_25))){ ambigs[1] = value; } //save value
if (((totalSoFar <= ptile25) && (totalSoFar > ptile0_25)) || ((lastValue < ptile25) && (totalSoFar > ptile25))) { ambigs[2] = value; } //save value
if (((totalSoFar <= ptile50) && (totalSoFar > ptile25)) || ((lastValue < ptile50) && (totalSoFar > ptile50))) { ambigs[3] = value; } //save value
if (((totalSoFar <= ptile75) && (totalSoFar > ptile50)) || ((lastValue < ptile75) && (totalSoFar > ptile75))) { ambigs[4] = value; } //save value
if (((totalSoFar <= ptile97_5) && (totalSoFar > ptile75)) || ((lastValue < ptile97_5) && (totalSoFar > ptile97_5))) { ambigs[5] = value; } //save value
if ((totalSoFar <= ptile100) && (totalSoFar > ptile97_5)) { ambigs[6] = value; } //save value
lastValue = totalSoFar;
}
ambigs[6] = (ambigBases.rbegin())->first;
if ((longHomoPolymer.begin())->first == -1) { homops[0] = 0; }
else {homops[0] = (longHomoPolymer.begin())->first; }
//set all values to min
homops[1] = homops[0]; homops[2] = homops[0]; homops[3] = homops[0]; homops[4] = homops[0]; homops[5] = homops[0];
totalSoFar = 0;
lastValue = 0;
for (map<int, long long>::iterator it = longHomoPolymer.begin(); it != longHomoPolymer.end(); it++) {
int value = it->first;
meanLongHomoPolymer += (it->first*it->second);
totalSoFar += it->second;
if (((totalSoFar <= ptile0_25) && (totalSoFar > 1)) || ((lastValue < ptile0_25) && (totalSoFar > ptile0_25))){ homops[1] = value; } //save value
if (((totalSoFar <= ptile25) && (totalSoFar > ptile0_25)) || ((lastValue < ptile25) && (totalSoFar > ptile25))) { homops[2] = value; } //save value
if (((totalSoFar <= ptile50) && (totalSoFar > ptile25)) || ((lastValue < ptile50) && (totalSoFar > ptile50))) { homops[3] = value; } //save value
if (((totalSoFar <= ptile75) && (totalSoFar > ptile50)) || ((lastValue < ptile75) && (totalSoFar > ptile75))) { homops[4] = value; } //save value
if (((totalSoFar <= ptile97_5) && (totalSoFar > ptile75)) || ((lastValue < ptile97_5) && (totalSoFar > ptile97_5))) { homops[5] = value; } //save value
if ((totalSoFar <= ptile100) && (totalSoFar > ptile97_5)) { homops[6] = value; } //save value
lastValue = totalSoFar;
}
homops[6] = (longHomoPolymer.rbegin())->first;
double meanstartPosition, meanendPosition, meanseqLength, meanambigBases, meanlongHomoPolymer;
meanstartPosition = meanStartPosition / (double) size; meanendPosition = meanEndPosition /(double) size; meanlongHomoPolymer = meanLongHomoPolymer / (double) size; meanseqLength = meanSeqLength / (double) size; meanambigBases = meanAmbigBases /(double) size;
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
m->mothurOutEndLine();
m->mothurOut("\t\tStart\tEnd\tNBases\tAmbigs\tPolymer\tNumSeqs"); m->mothurOutEndLine();
m->mothurOut("Minimum:\t" + toString(starts[0]) + "\t" + toString(ends[0]) + "\t" + toString(lengths[0]) + "\t" + toString(ambigs[0]) + "\t" + toString(homops[0]) + "\t" + toString(1)); m->mothurOutEndLine();
m->mothurOut("2.5%-tile:\t" + toString(starts[1]) + "\t" + toString(ends[1]) + "\t" + toString(lengths[1]) + "\t" + toString(ambigs[1]) + "\t" + toString(homops[1]) + "\t" + toString(ptile0_25)); m->mothurOutEndLine();
m->mothurOut("25%-tile:\t" + toString(starts[2]) + "\t" + toString(ends[2]) + "\t" + toString(lengths[2]) + "\t" + toString(ambigs[2]) + "\t" + toString(homops[2]) + "\t" + toString(ptile25)); m->mothurOutEndLine();
m->mothurOut("Median: \t" + toString(starts[3]) + "\t" + toString(ends[3]) + "\t" + toString(lengths[3]) + "\t" + toString(ambigs[3]) + "\t" + toString(homops[3]) + "\t" + toString(ptile50)); m->mothurOutEndLine();
m->mothurOut("75%-tile:\t" + toString(starts[4]) + "\t" + toString(ends[4]) + "\t" + toString(lengths[4]) + "\t" + toString(ambigs[4]) + "\t" + toString(homops[4]) + "\t" + toString(ptile75)); m->mothurOutEndLine();
m->mothurOut("97.5%-tile:\t" + toString(starts[5]) + "\t" + toString(ends[5]) + "\t" + toString(lengths[5]) + "\t" + toString(ambigs[5]) + "\t" + toString(homops[5]) + "\t" + toString(ptile97_5)); m->mothurOutEndLine();
m->mothurOut("Maximum:\t" + toString(starts[6]) + "\t" + toString(ends[6]) + "\t" + toString(lengths[6]) + "\t" + toString(ambigs[6]) + "\t" + toString(homops[6]) + "\t" + toString(ptile100)); m->mothurOutEndLine();
m->mothurOut("Mean:\t" + toString(meanstartPosition) + "\t" + toString(meanendPosition) + "\t" + toString(meanseqLength) + "\t" + toString(meanambigBases) + "\t" + toString(meanlongHomoPolymer)); m->mothurOutEndLine();
if ((namefile == "") && (countfile == "")) { m->mothurOut("# of Seqs:\t" + toString(numSeqs)); m->mothurOutEndLine(); }
else { m->mothurOut("# of unique seqs:\t" + toString(numSeqs)); m->mothurOutEndLine(); m->mothurOut("total # of seqs:\t" + toString(size)); m->mothurOutEndLine(); }
if (m->control_pressed) { m->mothurRemove(summaryFile); return 0; }
m->mothurOutEndLine();
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
m->mothurOut(summaryFile); m->mothurOutEndLine(); outputNames.push_back(summaryFile); outputTypes["summary"].push_back(summaryFile);
m->mothurOutEndLine();
if ((namefile == "") && (countfile == "")) { m->mothurOut("It took " + toString(time(NULL) - start) + " secs to summarize " + toString(numSeqs) + " sequences.\n"); }
else{ m->mothurOut("It took " + toString(time(NULL) - start) + " secs to summarize " + toString(size) + " sequences.\n"); }
//set fasta file as new current fastafile
string current = "";
itTypes = outputTypes.find("summary");
if (itTypes != outputTypes.end()) {
if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setSummaryFile(current); }
}
return 0;
}
catch(exception& e) {
m->errorOut(e, "SeqSummaryCommand", "execute");
exit(1);
}
}
