#include "filterseqscommand.h"
#include "sequence.hpp"
-/**************************************************************************************/
+//**********************************************************************************************************************
+vector<string> FilterSeqsCommand::setParameters(){
+ try {
+ CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta);
+ CommandParameter phard("hard", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(phard);
+ CommandParameter ptrump("trump", "String", "", "*", "", "", "",false,false); parameters.push_back(ptrump);
+ CommandParameter psoft("soft", "Number", "", "0", "", "", "",false,false); parameters.push_back(psoft);
+ CommandParameter pvertical("vertical", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pvertical);
+ CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
+ CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
+ CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
+
+ vector<string> myArray;
+ for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
+ return myArray;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "setParameters");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+string FilterSeqsCommand::getHelpString(){
+ try {
+ string helpString = "";
+ helpString += "The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n";
+ helpString += "The filter.seqs command parameters are fasta, trump, soft, hard, processors and vertical. \n";
+ helpString += "The fasta parameter is required, unless you have a valid current fasta file. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n";
+ helpString += "For example: fasta=abrecovery.fasta-amazon.fasta \n";
+ helpString += "The trump option will remove a column if the trump character is found at that position in any sequence of the alignment. Default=*, meaning no trump. \n";
+ helpString += "A soft mask removes any column where the dominant base (i.e. A, T, G, C, or U) does not occur in at least a designated percentage of sequences. Default=0.\n";
+ helpString += "The hard parameter allows you to enter a file containing the filter you want to use.\n";
+ helpString += "The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n";
+ helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n";
+ helpString += "The filter.seqs command should be in the following format: \n";
+ helpString += "filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n";
+ helpString += "Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n";
+ helpString += "Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n";
+ return helpString;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "getHelpString");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+FilterSeqsCommand::FilterSeqsCommand(){
+ try {
+ abort = true; calledHelp = true;
+ setParameters();
+ vector<string> tempOutNames;
+ outputTypes["fasta"] = tempOutNames;
+ outputTypes["filter"] = tempOutNames;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "FilterSeqsCommand");
+ exit(1);
+ }
+}
+/**************************************************************************************/
FilterSeqsCommand::FilterSeqsCommand(string option) {
try {
- abort = false;
+ abort = false; calledHelp = false;
filterFileName = "";
//allow user to run help
- if(option == "help") { help(); abort = true; }
+ if(option == "help") { help(); abort = true; calledHelp = true; }
+ else if(option == "citation") { citation(); abort = true; calledHelp = true;}
else {
- //valid paramters for this command
- string Array[] = {"fasta", "trump", "soft", "hard", "vertical", "outputdir","inputdir", "processors"};
- vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ vector<string> myArray = setParameters();
OptionParser parser(option);
map<string,string> parameters = parser.getParameters();
if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
}
+ //initialize outputTypes
+ vector<string> tempOutNames;
+ outputTypes["fasta"] = tempOutNames;
+ outputTypes["filter"] = 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 = ""; }
//check for required parameters
fasta = validParameter.validFile(parameters, "fasta", false);
- if (fasta == "not found") { m->mothurOut("fasta is a required parameter for the filter.seqs command."); m->mothurOutEndLine(); abort = true; }
+ if (fasta == "not found") {
+ fasta = m->getFastaFile();
+ if (fasta != "") { fastafileNames.push_back(fasta); m->mothurOut("Using " + fasta + " as input file for the fasta parameter."); m->mothurOutEndLine(); }
+ else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); m->mothurOutEndLine(); abort = true; }
+ }
else {
m->splitAtDash(fasta, fastafileNames);
//go through files and make sure they are good, if not, then disregard them
for (int i = 0; i < fastafileNames.size(); i++) {
- if (inputDir != "") {
- string path = m->hasPath(fastafileNames[i]);
- //if the user has not given a path then, add inputdir. else leave path alone.
- if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; }
- }
-
- ifstream in;
- int ableToOpen = m->openInputFile(fastafileNames[i], in, "noerror");
-
- //if you can't open it, try default location
- if (ableToOpen == 1) {
- if (m->getDefaultPath() != "") { //default path is set
- string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]);
- m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
- ableToOpen = m->openInputFile(tryPath, in, "noerror");
- fastafileNames[i] = tryPath;
+
+ bool ignore = false;
+ if (fastafileNames[i] == "current") {
+ fastafileNames[i] = m->getFastaFile();
+ if (fastafileNames[i] != "") { m->mothurOut("Using " + fastafileNames[i] + " as input file for the fasta parameter where you had given current."); m->mothurOutEndLine(); }
+ else {
+ m->mothurOut("You have no current fastafile, ignoring current."); m->mothurOutEndLine(); ignore=true;
+ //erase from file list
+ fastafileNames.erase(fastafileNames.begin()+i);
+ i--;
}
}
- in.close();
- if (ableToOpen == 1) {
- m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
- //erase from file list
- fastafileNames.erase(fastafileNames.begin()+i);
- i--;
- }else{
- string simpleName = m->getSimpleName(fastafileNames[i]);
- filterFileName += simpleName.substr(0, simpleName.find_first_of('.'));
+ if (!ignore) {
+ if (inputDir != "") {
+ string path = m->hasPath(fastafileNames[i]);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; }
+ }
+
+ ifstream in;
+ int ableToOpen = m->openInputFile(fastafileNames[i], in, "noerror");
+
+ //if you can't open it, try default location
+ if (ableToOpen == 1) {
+ if (m->getDefaultPath() != "") { //default path is set
+ string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]);
+ m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
+ ifstream in2;
+ ableToOpen = m->openInputFile(tryPath, in2, "noerror");
+ in2.close();
+ fastafileNames[i] = tryPath;
+ }
+ }
+
+ //if you can't open it, try default location
+ if (ableToOpen == 1) {
+ if (m->getOutputDir() != "") { //default path is set
+ string tryPath = m->getOutputDir() + m->getSimpleName(fastafileNames[i]);
+ m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
+ ifstream in2;
+ ableToOpen = m->openInputFile(tryPath, in2, "noerror");
+ in2.close();
+ fastafileNames[i] = tryPath;
+ }
+ }
+
+ in.close();
+
+ if (ableToOpen == 1) {
+ m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
+ //erase from file list
+ fastafileNames.erase(fastafileNames.begin()+i);
+ i--;
+ }else{
+ string simpleName = m->getSimpleName(fastafileNames[i]);
+ filterFileName += simpleName.substr(0, simpleName.find_first_of('.'));
+ m->setFastaFile(fastafileNames[i]);
+ }
+ in.close();
}
- in.close();
}
//make sure there is at least one valid file left
temp = validParameter.validFile(parameters, "soft", false); if (temp == "not found") { soft = 0; }
else { soft = (float)atoi(temp.c_str()) / 100.0; }
- temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
+ temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
+ m->setProcessors(temp);
convert(temp, processors);
vertical = validParameter.validFile(parameters, "vertical", false);
exit(1);
}
}
-
-//**********************************************************************************************************************
-
-void FilterSeqsCommand::help(){
- try {
-
- m->mothurOut("The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n");
- m->mothurOut("The filter.seqs command parameters are fasta, trump, soft, hard and vertical. \n");
- m->mothurOut("The fasta parameter is required. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n");
- m->mothurOut("For example: fasta=abrecovery.fasta-amazon.fasta \n");
- m->mothurOut("The trump parameter .... The default is ...\n");
- m->mothurOut("The soft parameter .... The default is ....\n");
- m->mothurOut("The hard parameter allows you to enter a file containing the filter you want to use.\n");
- m->mothurOut("The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n");
- m->mothurOut("The filter.seqs command should be in the following format: \n");
- m->mothurOut("filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n");
- m->mothurOut("Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n");
- m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n");
-
- }
- catch(exception& e) {
- m->errorOut(e, "FilterSeqsCommand", "help");
- exit(1);
- }
-}
-
/**************************************************************************************/
int FilterSeqsCommand::execute() {
try {
- if (abort == true) { return 0; }
+ if (abort == true) { if (calledHelp) { return 0; } return 2; }
ifstream inFASTA;
m->openInputFile(fastafileNames[0], inFASTA);
m->mothurOutEndLine(); m->mothurOutEndLine();
- if (m->control_pressed) { return 0; }
+ if (m->control_pressed) { outputTypes.clear(); return 0; }
#ifdef USE_MPI
int pid;
ofstream outFilter;
- string filterFile = outputDir + filterFileName + ".filter";
+ //prevent giantic file name
+ string filterFile;
+ if (fastafileNames.size() > 3) { filterFile = outputDir + "merge.filter"; }
+ else { filterFile = outputDir + filterFileName + ".filter"; }
+
m->openOutputFile(filterFile, outFilter);
outFilter << filter << endl;
outFilter.close();
- outputNames.push_back(filterFile);
+ outputNames.push_back(filterFile); outputTypes["filter"].push_back(filterFile);
#ifdef USE_MPI
}
if(filter[i] == '1'){ filteredLength++; }
}
- if (m->control_pressed) { for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+ if (m->control_pressed) { outputTypes.clear(); for(int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
m->mothurOutEndLine();
m->mothurOut("Length of the original alignment: " + toString(alignmentLength)); m->mothurOutEndLine();
m->mothurOut("Number of sequences used to construct filter: " + toString(numSeqs)); m->mothurOutEndLine();
+ //set fasta file as new current fastafile
+ string current = "";
+ itTypes = outputTypes.find("fasta");
+ if (itTypes != outputTypes.end()) {
+ if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setFastaFile(current); }
+ }
m->mothurOutEndLine();
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
string filteredFasta = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])) + "filter.fasta";
#ifdef USE_MPI
- int pid, start, end, numSeqsPerProcessor, num;
+ int pid, numSeqsPerProcessor, num;
int tag = 2001;
- vector<unsigned long int>MPIPos;
+ vector<unsigned long long>MPIPos;
MPI_Status status;
MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
MPI_File outMPI;
- MPI_File tempMPI;
MPI_File inMPI;
int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
int inMode=MPI_MODE_RDONLY;
//wait on chidren
for(int i = 1; i < processors; i++) {
- 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
if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); 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);
MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
#else
- vector<unsigned long int> positions = m->divideFile(fastafileNames[s], processors);
+ vector<unsigned long long> positions = m->divideFile(fastafileNames[s], processors);
for (int i = 0; i < (positions.size()-1); i++) {
lines.push_back(new linePair(positions[i], positions[(i+1)]));
//append fasta files
for(int i=1;i<processors;i++){
m->appendFiles((fastafileNames[s] + toString(processIDS[i]) + ".temp"), filteredFasta);
- remove((fastafileNames[s] + toString(processIDS[i]) + ".temp").c_str());
+ m->mothurRemove((fastafileNames[s] + toString(processIDS[i]) + ".temp"));
}
}
if (m->control_pressed) { return 1; }
#endif
#endif
- outputNames.push_back(filteredFasta);
+ outputNames.push_back(filteredFasta); outputTypes["fasta"].push_back(filteredFasta);
}
return 0;
}
#ifdef USE_MPI
/**************************************************************************************/
-int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector<unsigned long int>& MPIPos) {
+int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector<unsigned long long>& MPIPos) {
try {
string outputString = "";
int count = 0;
}
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- unsigned long int pos = in.tellg();
+ unsigned long long pos = in.tellg();
if ((pos == -1) || (pos >= filePos->end)) { break; }
#else
if (in.eof()) { break; }
out.close();
exit(0);
- }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); 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);
+ }
}
//force parent to wait until all the processes are done
string tempFile = filename + toString(processIDS[i]) + ".num.temp";
m->openInputFile(tempFile, in);
if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
- in.close(); remove(tempFile.c_str());
+ in.close(); m->mothurRemove(tempFile);
}
#ifdef USE_MPI
int pid, numSeqsPerProcessor, num;
int tag = 2001;
- vector<unsigned long int> MPIPos;
+ vector<unsigned long long> MPIPos;
MPI_Status status;
MPI_File inMPI;
MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
#else
- vector<unsigned long int> positions = m->divideFile(fastafileNames[s], processors);
+ vector<unsigned long long> positions = m->divideFile(fastafileNames[s], processors);
for (int i = 0; i < (positions.size()-1); i++) {
lines.push_back(new linePair(positions[i], positions[(i+1)]));
}
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- unsigned long int pos = in.tellg();
+ unsigned long long pos = in.tellg();
if ((pos == -1) || (pos >= filePos->end)) { break; }
#else
if (in.eof()) { break; }
}
#ifdef USE_MPI
/**************************************************************************************/
-int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector<unsigned long int>& MPIPos) {
+int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector<unsigned long long>& MPIPos) {
try {
MPI_Status status;
int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) {
try {
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- int process = 0;
+ int process = 1;
int num = 0;
processIDS.clear();
out.close();
exit(0);
- }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); 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);
+ }
}
+ //parent do your part
+ num = driverCreateFilter(F, filename, lines[0]);
+
//force parent to wait until all the processes are done
- for (int i=0;i<processors;i++) {
+ for (int i=0;i<(processors-1);i++) {
int temp = processIDS[i];
wait(&temp);
}
- //parent reads in and combine Filter info
+ //parent reads in and combines Filter info
for (int i = 0; i < processIDS.size(); i++) {
string tempFilename = filename + toString(processIDS[i]) + "filterValues.temp";
ifstream in;
for (int k = 0; k < alignmentLength; k++) { in >> temp; F.gap[k] += temp; } m->gobble(in);
in.close();
- remove(tempFilename.c_str());
+ m->mothurRemove(tempFilename);
}
return num;
projects / mothur.git / blobdiff