#include "chimerauchimecommand.h"
#include "deconvolutecommand.h"
-#include "uc.h"
+//#include "uc.h"
#include "sequence.hpp"
#include "referencedb.h"
-
+#include "systemcommand.h"
//**********************************************************************************************************************
vector<string> ChimeraUchimeCommand::setParameters(){
}
}
//**********************************************************************************************************************
+string ChimeraUchimeCommand::getOutputFileNameTag(string type, string inputName=""){
+ try {
+ string outputFileName = "";
+ map<string, vector<string> >::iterator it;
+
+ //is this a type this command creates
+ it = outputTypes.find(type);
+ if (it == outputTypes.end()) { m->mothurOut("[ERROR]: this command doesn't create a " + type + " output file.\n"); }
+ else {
+ if (type == "chimera") { outputFileName = "uchime.chimeras"; }
+ else if (type == "accnos") { outputFileName = "uchime.accnos"; }
+ else if (type == "alns") { outputFileName = "uchime.alns"; }
+ else { m->mothurOut("[ERROR]: No definition for type " + type + " output file tag.\n"); m->control_pressed = true; }
+ }
+ return outputFileName;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ChimeraUchimeCommand", "getOutputFileNameTag");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
ChimeraUchimeCommand::ChimeraUchimeCommand(){
try {
abort = true; calledHelp = true;
string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
m->setProcessors(temp);
- convert(temp, processors);
+ m->mothurConvert(temp, processors);
abskew = validParameter.validFile(parameters, "abskew", false); if (abskew == "not found"){ useAbskew = false; abskew = "1.9"; }else{ useAbskew = true; }
if (useAbskew && templatefile != "self") { m->mothurOut("The abskew parameter is only valid with template=self, ignoring."); m->mothurOutEndLine(); useAbskew = false; }
if (hasName && (templatefile != "self")) { m->mothurOut("You have provided a namefile and the reference parameter is not set to self. I am not sure what reference you are trying to use, aborting."); m->mothurOutEndLine(); abort=true; }
if (hasGroup && (templatefile != "self")) { m->mothurOut("You have provided a group file and the reference parameter is not set to self. I am not sure what reference you are trying to use, aborting."); m->mothurOutEndLine(); abort=true; }
+
+ //look for uchime exe
+ path = m->argv;
+ string tempPath = path;
+ for (int i = 0; i < path.length(); i++) { tempPath[i] = tolower(path[i]); }
+ path = path.substr(0, (tempPath.find_last_of('m')));
+
+ string uchimeCommand;
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+ uchimeCommand = path + "uchime"; // format the database, -o option gives us the ability
+ if (m->debug) {
+ m->mothurOut("[DEBUG]: Uchime location using \"which uchime\" = ");
+ Command* newCommand = new SystemCommand("which uchime"); m->mothurOutEndLine();
+ newCommand->execute();
+ delete newCommand;
+ m->mothurOut("[DEBUG]: Mothur's location using \"which mothur\" = ");
+ newCommand = new SystemCommand("which mothur"); m->mothurOutEndLine();
+ newCommand->execute();
+ delete newCommand;
+ }
+#else
+ uchimeCommand = path + "uchime.exe";
+#endif
+
+ //test to make sure uchime exists
+ ifstream in;
+ uchimeCommand = m->getFullPathName(uchimeCommand);
+ int ableToOpen = m->openInputFile(uchimeCommand, in, "no error"); in.close();
+ if(ableToOpen == 1) {
+ m->mothurOut(uchimeCommand + " file does not exist. Checking path... \n");
+ //check to see if uchime is in the path??
+
+ string uLocation = m->findProgramPath("uchime");
+
+
+ ifstream in2;
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+ ableToOpen = m->openInputFile(uLocation, in2, "no error"); in2.close();
+#else
+ ableToOpen = m->openInputFile((uLocation + ".exe"), in2, "no error"); in2.close();
+#endif
- }
+ if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + uLocation + " file does not exist. mothur requires the uchime executable."); m->mothurOutEndLine(); abort = true; }
+ else { m->mothurOut("Found uchime in your path, using " + uLocation + "\n");uchimeLocation = uLocation; }
+ }else { uchimeLocation = uchimeCommand; }
+
+ uchimeLocation = m->getFullPathName(uchimeLocation);
+ }
}
catch(exception& e) {
m->errorOut(e, "ChimeraSlayerCommand", "ChimeraSlayerCommand");
int start = time(NULL);
string nameFile = "";
if (outputDir == "") { outputDir = m->hasPath(fastaFileNames[s]); }//if user entered a file with a path then preserve it
- string outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "uchime.chimera";
- string accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "uchime.accnos";
- string alnsFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "uchime.alns";
+ string outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + getOutputFileNameTag("chimera");
+ string accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + getOutputFileNameTag("accnos");
+ string alnsFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + getOutputFileNameTag("alns");
string newFasta = m->getRootName(fastaFileNames[s]) + "temp";
//you provided a groupfile
if (nameFileNames.size() != 0) { //you provided a namefile and we don't need to create one
nameFile = nameFileNames[s];
}else { nameFile = getNamesFile(fastaFileNames[s]); }
-
+
map<string, string> seqs;
readFasta(fastaFileNames[s], seqs); if (m->control_pressed) { for (int j = 0; j < outputNames.size(); j++) { m->mothurRemove(outputNames[j]); } return 0; }
if (chimealns) { m->openOutputFile(alnsFileName, out2); out2.close(); }
int totalSeqs = 0;
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
if(processors == 1) { totalSeqs = driverGroups(parser, outputFileName, newFasta, accnosFileName, alnsFileName, 0, groups.size(), groups); }
- else { totalSeqs = createProcessesGroups(parser, outputFileName, newFasta, accnosFileName, alnsFileName, groups); }
- #else
- totalSeqs = driverGroups(parser, outputFileName, newFasta, accnosFileName, alnsFileName, 0, groups.size(), groups);
- #endif
+ else { totalSeqs = createProcessesGroups(parser, outputFileName, newFasta, accnosFileName, alnsFileName, groups, nameFile, groupFile, fastaFileNames[s]); }
+
if (m->control_pressed) { for (int j = 0; j < outputNames.size(); j++) { m->mothurRemove(outputNames[j]); } return 0; }
int totalChimeras = deconvoluteResults(parser, outputFileName, accnosFileName, alnsFileName);
int numSeqs = 0;
int numChimeras = 0;
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+
if(processors == 1){ numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName, numChimeras); }
else{ numSeqs = createProcesses(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName, numChimeras); }
- #else
- numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName, numChimeras);
- #endif
+
+ //add headings
+ ofstream out;
+ m->openOutputFile(outputFileName+".temp", out);
+ out << "Score\tQuery\tParentA\tParentB\tIdQM\tIdQA\tIdQB\tIdAB\tIdQT\tLY\tLN\tLA\tRY\tRN\tRA\tDiv\tYN\n";
+ out.close();
+
+ m->appendFiles(outputFileName, outputFileName+".temp");
+ m->mothurRemove(outputFileName); rename((outputFileName+".temp").c_str(), outputFileName.c_str());
+
if (m->control_pressed) { for (int j = 0; j < outputNames.size(); j++) { m->mothurRemove(outputNames[j]); } return 0; }
//remove file made for uchime
map<string, string>::iterator itUnique;
int total = 0;
+ //edit accnos file
+ ifstream in2;
+ m->openInputFile(accnosFileName, in2);
+
+ ofstream out2;
+ m->openOutputFile(accnosFileName+".temp", out2);
+
+ string name;
+ set<string> namesInFile; //this is so if a sequence is found to be chimera in several samples we dont write it to the results file more than once
+ set<string>::iterator itNames;
+ set<string> chimerasInFile;
+ set<string>::iterator itChimeras;
+
+
+ while (!in2.eof()) {
+ if (m->control_pressed) { in2.close(); out2.close(); m->mothurRemove(outputFileName); m->mothurRemove((accnosFileName+".temp")); return 0; }
+
+ in2 >> name; m->gobble(in2);
+
+ //find unique name
+ itUnique = uniqueNames.find(name);
+
+ if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing accnos results. Cannot find "+ name + "."); m->mothurOutEndLine(); m->control_pressed = true; }
+ else {
+ itChimeras = chimerasInFile.find((itUnique->second));
+
+ if (itChimeras == chimerasInFile.end()) {
+ out2 << itUnique->second << endl;
+ chimerasInFile.insert((itUnique->second));
+ total++;
+ }
+ }
+ }
+ in2.close();
+ out2.close();
+
+ m->mothurRemove(accnosFileName);
+ rename((accnosFileName+".temp").c_str(), accnosFileName.c_str());
+
+
+
//edit chimera file
ifstream in;
m->openInputFile(outputFileName, in);
ofstream out;
m->openOutputFile(outputFileName+".temp", out); out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
+ out << "Score\tQuery\tParentA\tParentB\tIdQM\tIdQA\tIdQB\tIdAB\tIdQT\tLY\tLN\tLA\tRY\tRN\tRA\tDiv\tYN\n";
float temp1;
- string name, rest, parent1, parent2;
- set<string> namesInFile; //this is so if a sequence is found to be chimera in several samples we dont write it to the results file more than once
- set<string>::iterator itNames;
-
+ string parent1, parent2, temp2, temp3, temp4, temp5, temp6, temp7, temp8, temp9, temp10, temp11, temp12, temp13, flag;
+ name = "";
+ namesInFile.clear();
//assumptions - in file each read will always look like - if uchime source is updated, revisit this code.
- /*
+ /* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0.000000 F11Fcsw_33372/ab=18/ * * * * * * * * * * * * * * N
0.018300 F11Fcsw_14980/ab=16/ F11Fcsw_1915/ab=35/ F11Fcsw_6032/ab=42/ 79.9 78.7 78.2 78.7 79.2 3 0 5 11 10 20 1.46 N
*/
if (m->control_pressed) { in.close(); out.close(); m->mothurRemove((outputFileName+".temp")); return 0; }
+ bool print = false;
in >> temp1; m->gobble(in);
in >> name; m->gobble(in);
in >> parent1; m->gobble(in);
in >> parent2; m->gobble(in);
- rest = m->getline(in); m->gobble(in);
+ in >> temp2 >> temp3 >> temp4 >> temp5 >> temp6 >> temp7 >> temp8 >> temp9 >> temp10 >> temp11 >> temp12 >> temp13 >> flag;
+ m->gobble(in);
//parse name - name will look like U68590/ab=1/
string restOfName = "";
if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing chimera results. Cannot find "+ name + "."); m->mothurOutEndLine(); m->control_pressed = true; }
else {
- itNames = namesInFile.find((itUnique->second));
+ name = itUnique->second;
+ //is this name already in the file
+ itNames = namesInFile.find((name));
- if (itNames == namesInFile.end()) {
- out << temp1 << '\t' << itUnique->second << restOfName << '\t';
- namesInFile.insert((itUnique->second));
-
- //parse parent1 names
- if (parent1 != "*") {
- restOfName = "";
- pos = parent1.find_first_of('/');
- if (pos != string::npos) {
- restOfName = parent1.substr(pos);
- parent1 = parent1.substr(0, pos);
- }
+ if (itNames == namesInFile.end()) { //no not in file
+ if (flag == "N") { //are you really a no??
+ //is this sequence really not chimeric??
+ itChimeras = chimerasInFile.find(name);
- itUnique = uniqueNames.find(parent1);
- if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing chimera results. Cannot find parentA "+ parent1 + "."); m->mothurOutEndLine(); m->control_pressed = true; }
- else {
- out << itUnique->second << restOfName << '\t';
- }
- }else { out << parent1 << '\t'; }
+ //then you really are a no so print, otherwise skip
+ if (itChimeras == chimerasInFile.end()) { print = true; }
+ }else{ print = true; }
+ }
+ }
+
+ if (print) {
+ out << temp1 << '\t' << name << restOfName << '\t';
+ namesInFile.insert(name);
+
+ //parse parent1 names
+ if (parent1 != "*") {
+ restOfName = "";
+ pos = parent1.find_first_of('/');
+ if (pos != string::npos) {
+ restOfName = parent1.substr(pos);
+ parent1 = parent1.substr(0, pos);
+ }
- //parse parent2 names
- if (parent2 != "*") {
- restOfName = "";
- pos = parent2.find_first_of('/');
- if (pos != string::npos) {
- restOfName = parent2.substr(pos);
- parent2 = parent2.substr(0, pos);
- }
-
- itUnique = uniqueNames.find(parent2);
- if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing chimera results. Cannot find parentB "+ parent2 + "."); m->mothurOutEndLine(); m->control_pressed = true; }
- else {
- out << itUnique->second << restOfName << '\t';
- }
- }else { out << parent2 << '\t'; }
+ itUnique = uniqueNames.find(parent1);
+ if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing chimera results. Cannot find parentA "+ parent1 + "."); m->mothurOutEndLine(); m->control_pressed = true; }
+ else { out << itUnique->second << restOfName << '\t'; }
+ }else { out << parent1 << '\t'; }
+
+ //parse parent2 names
+ if (parent2 != "*") {
+ restOfName = "";
+ pos = parent2.find_first_of('/');
+ if (pos != string::npos) {
+ restOfName = parent2.substr(pos);
+ parent2 = parent2.substr(0, pos);
+ }
- out << rest << endl;
- }
+ itUnique = uniqueNames.find(parent2);
+ if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing chimera results. Cannot find parentB "+ parent2 + "."); m->mothurOutEndLine(); m->control_pressed = true; }
+ else { out << itUnique->second << restOfName << '\t'; }
+ }else { out << parent2 << '\t'; }
+
+ out << temp2 << '\t' << temp3 << '\t' << temp4 << '\t' << temp5 << '\t' << temp6 << '\t' << temp7 << '\t' << temp8 << '\t' << temp9 << '\t' << temp10 << '\t' << temp11 << '\t' << temp12 << temp13 << '\t' << flag << endl;
}
}
in.close();
m->mothurRemove(outputFileName);
rename((outputFileName+".temp").c_str(), outputFileName.c_str());
- //edit accnos file
- ifstream in2;
- m->openInputFile(accnosFileName, in2);
-
- ofstream out2;
- m->openOutputFile(accnosFileName+".temp", out2);
-
- name = "";
- namesInFile.clear();
-
- while (!in2.eof()) {
- if (m->control_pressed) { in2.close(); out2.close(); m->mothurRemove(outputFileName); m->mothurRemove((accnosFileName+".temp")); return 0; }
-
- in2 >> name; m->gobble(in2);
-
- //find unique name
- itUnique = uniqueNames.find(name);
-
- if (itUnique == uniqueNames.end()) { m->mothurOut("[ERROR]: trouble parsing accnos results. Cannot find "+ name + "."); m->mothurOutEndLine(); m->control_pressed = true; }
- else {
- itNames = namesInFile.find((itUnique->second));
-
- if (itNames == namesInFile.end()) {
- out2 << itUnique->second << endl;
- namesInFile.insert((itUnique->second));
- total++;
- }
- }
- }
- in2.close();
- out2.close();
-
- m->mothurRemove(accnosFileName);
- rename((accnosFileName+".temp").c_str(), accnosFileName.c_str());
-
+
//edit anls file
//assumptions - in file each read will always look like - if uchime source is updated, revisit this code.
/*
string inputString = "fasta=" + inputFile;
m->mothurOut("/******************************************/"); m->mothurOutEndLine();
m->mothurOut("Running command: unique.seqs(" + inputString + ")"); m->mothurOutEndLine();
-
+ m->mothurCalling = true;
+
Command* uniqueCommand = new DeconvoluteCommand(inputString);
uniqueCommand->execute();
map<string, vector<string> > filenames = uniqueCommand->getOutputFiles();
delete uniqueCommand;
-
+ m->mothurCalling = false;
m->mothurOut("/******************************************/"); m->mothurOutEndLine();
nameFile = filenames["name"][0];
if (m->control_pressed) { return 0; }
//remove file made for uchime
- m->mothurRemove(filename);
+ if (!m->debug) { m->mothurRemove(filename); }
+ else { m->mothurOut("[DEBUG]: saving file: " + filename + ".\n"); }
//append files
m->appendFiles((outputFName+groups[i]), outputFName); m->mothurRemove((outputFName+groups[i]));
int ChimeraUchimeCommand::driver(string outputFName, string filename, string accnos, string alns, int& numChimeras){
try {
+ outputFName = m->getFullPathName(outputFName);
+ filename = m->getFullPathName(filename);
+ alns = m->getFullPathName(alns);
+
+ //to allow for spaces in the path
+ outputFName = "\"" + outputFName + "\"";
+ filename = "\"" + filename + "\"";
+ alns = "\"" + alns + "\"";
+
vector<char*> cPara;
- char* tempUchime = new char[8];
- strcpy(tempUchime, "./uchime ");
+ string uchimeCommand = uchimeLocation;
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+ uchimeCommand += " ";
+#else
+ uchimeCommand = "\"" + uchimeCommand + "\"";
+#endif
+
+ char* tempUchime;
+ tempUchime= new char[uchimeCommand.length()+1];
+ *tempUchime = '\0';
+ strncat(tempUchime, uchimeCommand.c_str(), uchimeCommand.length());
cPara.push_back(tempUchime);
char* tempIn = new char[8];
char** uchimeParameters;
uchimeParameters = new char*[cPara.size()];
- for (int i = 0; i < cPara.size(); i++) { uchimeParameters[i] = cPara[i]; }
- int numArgs = cPara.size();
-
- uchime_main(numArgs, uchimeParameters);
+ string commandString = "";
+ for (int i = 0; i < cPara.size(); i++) { uchimeParameters[i] = cPara[i]; commandString += toString(cPara[i]) + " "; }
+ //int numArgs = cPara.size();
+
+ //uchime_main(numArgs, uchimeParameters);
+ //cout << "commandString = " << commandString << endl;
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+#else
+ commandString = "\"" + commandString + "\"";
+#endif
+ if (m->debug) { m->mothurOut("[DEBUG]: uchime command = " + commandString + ".\n"); }
+ system(commandString.c_str());
//free memory
- for(int i = 0; i < cPara.size(); i++) { delete[] cPara[i]; }
+ for(int i = 0; i < cPara.size(); i++) { delete cPara[i]; }
delete[] uchimeParameters;
+ //remove "" from filenames
+ outputFName = outputFName.substr(1, outputFName.length()-2);
+ filename = filename.substr(1, filename.length()-2);
+ alns = alns.substr(1, alns.length()-2);
+
if (m->control_pressed) { return 0; }
//create accnos file from uchime results
processIDS.clear();
int process = 1;
int num = 0;
-#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- //break up file into multiple files
vector<string> files;
+
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+ //break up file into multiple files
m->divideFile(filename, processors, files);
if (m->control_pressed) { return 0; }
}
in.close(); m->mothurRemove(tempFile);
}
+#else
+ //////////////////////////////////////////////////////////////////////////////////////////////////////
+ //Windows version shared memory, so be careful when passing variables through the preClusterData struct.
+ //Above fork() will clone, so memory is separate, but that's not the case with windows,
+ //////////////////////////////////////////////////////////////////////////////////////////////////////
+
+ //divide file
+ int count = 0;
+ int spot = 0;
+ map<int, ofstream*> filehandles;
+ map<int, ofstream*>::iterator it3;
+
+ ofstream* temp;
+ for (int i = 0; i < processors; i++) {
+ temp = new ofstream;
+ filehandles[i] = temp;
+ m->openOutputFile(filename+toString(i)+".temp", *(temp));
+ files.push_back(filename+toString(i)+".temp");
+ }
+
+ ifstream in;
+ m->openInputFile(filename, in);
+
+ while(!in.eof()) {
+
+ if (m->control_pressed) { in.close(); for (it3 = filehandles.begin(); it3 != filehandles.end(); it3++) { (*(it3->second)).close(); delete it3->second; } return 0; }
+
+ Sequence tempSeq(in); m->gobble(in);
+
+ if (tempSeq.getName() != "") {
+ tempSeq.printSequence(*(filehandles[spot]));
+ spot++; count++;
+ if (spot == processors) { spot = 0; }
+ }
+ }
+ in.close();
+
+ //delete memory
+ for (it3 = filehandles.begin(); it3 != filehandles.end(); it3++) {
+ (*(it3->second)).close();
+ delete it3->second;
+ }
+
+ //sanity check for number of processors
+ if (count < processors) { processors = count; }
+
+ vector<uchimeData*> pDataArray;
+ DWORD dwThreadIdArray[processors-1];
+ HANDLE hThreadArray[processors-1];
+ vector<string> dummy; //used so that we can use the same struct for MyUchimeSeqsThreadFunction and MyUchimeThreadFunction
+
+ //Create processor worker threads.
+ for( int i=1; i<processors; i++ ){
+ // Allocate memory for thread data.
+ string extension = toString(i) + ".temp";
+
+ uchimeData* tempUchime = new uchimeData(outputFileName+extension, uchimeLocation, templatefile, files[i], "", "", "", accnos+extension, alns+extension, dummy, m, 0, 0, i);
+ tempUchime->setBooleans(useAbskew, chimealns, useMinH, useMindiv, useXn, useDn, useXa, useChunks, useMinchunk, useIdsmoothwindow, useMinsmoothid, useMaxp, skipgaps, skipgaps2, useMinlen, useMaxlen, ucl, useQueryfract);
+ tempUchime->setVariables(abskew, minh, mindiv, xn, dn, xa, chunks, minchunk, idsmoothwindow, minsmoothid, maxp, minlen, maxlen, queryfract);
+
+ pDataArray.push_back(tempUchime);
+ processIDS.push_back(i);
+
+ //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-1] = CreateThread(NULL, 0, MyUchimeSeqsThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
+ }
+ //using the main process as a worker saves time and memory
+ num = driver(outputFileName, files[0], accnos, alns, numChimeras);
+
+ //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;
+ numChimeras += pDataArray[i]->numChimeras;
+ CloseHandle(hThreadArray[i]);
+ delete pDataArray[i];
+ }
+#endif
+
//append output files
- for(int i=0;i<processIDS[i];i++){
+ for(int i=0;i<processIDS.size();i++){
m->appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
m->mothurRemove((outputFileName + toString(processIDS[i]) + ".temp"));
//get rid of the file pieces.
for (int i = 0; i < files.size(); i++) { m->mothurRemove(files[i]); }
-#endif
return num;
}
catch(exception& e) {
}
/**************************************************************************************************/
-int ChimeraUchimeCommand::createProcessesGroups(SequenceParser& parser, string outputFName, string filename, string accnos, string alns, vector<string> groups) {
+int ChimeraUchimeCommand::createProcessesGroups(SequenceParser& parser, string outputFName, string filename, string accnos, string alns, vector<string> groups, string nameFile, string groupFile, string fastaFile) {
try {
processIDS.clear();
lines.push_back(linePair(startIndex, endIndex));
}
-#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
//loop through and create all the processes you want
while (process != processors) {
int temp = processIDS[i];
wait(&temp);
}
-#endif
for (int i = 0; i < processIDS.size(); i++) {
ifstream in;
if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
in.close(); m->mothurRemove(tempFile);
}
+
+#else
+ //////////////////////////////////////////////////////////////////////////////////////////////////////
+ //Windows version shared memory, so be careful when passing variables through the uchimeData struct.
+ //Above fork() will clone, so memory is separate, but that's not the case with windows,
+ //////////////////////////////////////////////////////////////////////////////////////////////////////
+
+ vector<uchimeData*> pDataArray;
+ DWORD dwThreadIdArray[processors-1];
+ HANDLE hThreadArray[processors-1];
+
+ //Create processor worker threads.
+ for( int i=1; i<processors; i++ ){
+ // Allocate memory for thread data.
+ string extension = toString(i) + ".temp";
+
+ uchimeData* tempUchime = new uchimeData(outputFName+extension, uchimeLocation, templatefile, filename+extension, fastaFile, nameFile, groupFile, accnos+extension, alns+extension, groups, m, lines[i].start, lines[i].end, i);
+ tempUchime->setBooleans(useAbskew, chimealns, useMinH, useMindiv, useXn, useDn, useXa, useChunks, useMinchunk, useIdsmoothwindow, useMinsmoothid, useMaxp, skipgaps, skipgaps2, useMinlen, useMaxlen, ucl, useQueryfract);
+ tempUchime->setVariables(abskew, minh, mindiv, xn, dn, xa, chunks, minchunk, idsmoothwindow, minsmoothid, maxp, minlen, maxlen, queryfract);
+
+ pDataArray.push_back(tempUchime);
+ processIDS.push_back(i);
+
+ //MyUchimeThreadFunction 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-1] = CreateThread(NULL, 0, MyUchimeThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
+ }
+
+
+ //using the main process as a worker saves time and memory
+ num = driverGroups(parser, outputFName, filename, accnos, alns, lines[0].start, lines[0].end, groups);
+ //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;
+ CloseHandle(hThreadArray[i]);
+ delete pDataArray[i];
+ }
+#endif
+
+
//append output files
- for(int i=0;i<processIDS[i];i++){
+ for(int i=0;i<processIDS.size();i++){
m->appendFiles((outputFName + toString(processIDS[i]) + ".temp"), outputFName);
m->mothurRemove((outputFName + toString(processIDS[i]) + ".temp"));
projects / mothur.git / blobdiff