X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=nmdscommand.cpp;h=a90ed2965888928f4a90049fe04fe888d773c450;hp=0661f24b09d4494d658a6044fdce1fafe0b5582f;hb=df7e3ff9f68ef157b0328a2d353c3258c5d45d89;hpb=6d12ed0ba66fb35e9e2781fe3ca361e2293f2476 diff --git a/nmdscommand.cpp b/nmdscommand.cpp index 0661f24..a90ed29 100644 --- a/nmdscommand.cpp +++ b/nmdscommand.cpp @@ -11,52 +11,79 @@ #include "readphylipvector.h" //********************************************************************************************************************** -vector NMDSCommand::getValidParameters(){ +vector NMDSCommand::setParameters(){ try { - string Array[] = {"phylip","axes","mindim","maxdim","iters","maxiters","epsilon","outputdir","inputdir"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + CommandParameter paxes("axes", "InputTypes", "", "", "none", "none", "none","",false,false,true); parameters.push_back(paxes); + CommandParameter pphylip("phylip", "InputTypes", "", "", "none", "none", "none","nmds-stress",false,true,true); parameters.push_back(pphylip); + CommandParameter pmaxdim("maxdim", "Number", "", "2", "", "", "","",false,false); parameters.push_back(pmaxdim); + CommandParameter pmindim("mindim", "Number", "", "2", "", "", "","",false,false); parameters.push_back(pmindim); + CommandParameter piters("iters", "Number", "", "10", "", "", "","",false,false); parameters.push_back(piters); + CommandParameter pmaxiters("maxiters", "Number", "", "500", "", "", "","",false,false); parameters.push_back(pmaxiters); + CommandParameter pepsilon("epsilon", "Number", "", "0.000000000001", "", "", "","",false,false); parameters.push_back(pepsilon); + CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir); + CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir); + + vector myArray; + for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); } return myArray; } catch(exception& e) { - m->errorOut(e, "NMDSCommand", "getValidParameters"); + m->errorOut(e, "NMDSCommand", "setParameters"); exit(1); } } //********************************************************************************************************************** -NMDSCommand::NMDSCommand(){ +string NMDSCommand::getHelpString(){ try { - abort = true; - //initialize outputTypes - vector tempOutNames; - outputTypes["nmds"] = tempOutNames; - outputTypes["stress"] = tempOutNames; - outputTypes["iters"] = tempOutNames; + string helpString = ""; + helpString += "The nmds command is modelled after the nmds code written in R by Sarah Goslee, using Non-metric multidimensional scaling function using the majorization algorithm from Borg & Groenen 1997, Modern Multidimensional Scaling.\n"; + helpString += "The nmds command parameters are phylip, axes, mindim, maxdim, maxiters, iters and epsilon.\n"; + helpString += "The phylip parameter allows you to enter your distance file.\n"; + helpString += "The axes parameter allows you to enter a file containing a starting configuration.\n"; + helpString += "The maxdim parameter allows you to select the maximum dimensions to use. Default=2\n"; + helpString += "The mindim parameter allows you to select the minimum dimensions to use. Default=2\n"; + helpString += "The maxiters parameter allows you to select the maximum number of iters to try with each random configuration. Default=500\n"; + helpString += "The iters parameter allows you to select the number of random configuration to try. Default=10\n"; + helpString += "The epsilon parameter allows you to select set an acceptable stopping point. Default=1e-12.\n"; + helpString += "Example nmds(phylip=yourDistanceFile).\n"; + helpString += "Note: No spaces between parameter labels (i.e. phylip), '=' and parameters (i.e.yourDistanceFile).\n"; + return helpString; } catch(exception& e) { - m->errorOut(e, "NMDSCommand", "NMDSCommand"); + m->errorOut(e, "NMDSCommand", "getHelpString"); exit(1); } } //********************************************************************************************************************** -vector NMDSCommand::getRequiredParameters(){ - try { - string Array[] = {"phylip"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); - return myArray; - } - catch(exception& e) { - m->errorOut(e, "NMDSCommand", "getRequiredParameters"); - exit(1); - } +string NMDSCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "nmds") { pattern = "[filename],nmds.axes"; } + else if (type == "stress") { pattern = "[filename],nmds.stress"; } + else if (type == "iters") { pattern = "[filename],nmds.iters"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "NMDSCommand", "getOutputPattern"); + exit(1); + } } + //********************************************************************************************************************** -vector NMDSCommand::getRequiredFiles(){ +NMDSCommand::NMDSCommand(){ try { - vector myArray; - return myArray; + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["nmds"] = tempOutNames; + outputTypes["stress"] = tempOutNames; + outputTypes["iters"] = tempOutNames; } catch(exception& e) { - m->errorOut(e, "NMDSCommand", "getRequiredFiles"); + m->errorOut(e, "NMDSCommand", "NMDSCommand"); exit(1); } } @@ -64,15 +91,14 @@ vector NMDSCommand::getRequiredFiles(){ NMDSCommand::NMDSCommand(string option) { try { - abort = false; + abort = false; calledHelp = false; //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[] = {"phylip","axes","mindim","maxdim","iters","maxiters","epsilon","outputdir", "inputdir"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser. getParameters(); @@ -115,7 +141,12 @@ NMDSCommand::NMDSCommand(string option) { //required parameters phylipfile = validParameter.validFile(parameters, "phylip", true); if (phylipfile == "not open") { phylipfile = ""; abort = true; } - else if (phylipfile == "not found") { phylipfile = ""; m->mothurOut("You must provide a distance file before running the nmds command."); m->mothurOutEndLine(); abort = true; } + else if (phylipfile == "not found") { + //if there is a current phylip file, use it + phylipfile = m->getPhylipFile(); + if (phylipfile != "") { m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); } + else { m->mothurOut("You have no current phylip file and the phylip parameter is required."); m->mothurOutEndLine(); abort = true; } + }else { m->setPhylipFile(phylipfile); } axesfile = validParameter.validFile(parameters, "axes", true); if (axesfile == "not open") { axesfile = ""; abort = true; } @@ -128,22 +159,22 @@ NMDSCommand::NMDSCommand(string option) { } string temp = validParameter.validFile(parameters, "mindim", false); if (temp == "not found") { temp = "2"; } - convert(temp, mindim); + m->mothurConvert(temp, mindim); temp = validParameter.validFile(parameters, "maxiters", false); if (temp == "not found") { temp = "500"; } - convert(temp, maxIters); + m->mothurConvert(temp, maxIters); temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "10"; } - convert(temp, iters); + m->mothurConvert(temp, iters); temp = validParameter.validFile(parameters, "maxdim", false); if (temp == "not found") { temp = "2"; } - convert(temp, maxdim); + m->mothurConvert(temp, maxdim); temp = validParameter.validFile(parameters, "epsilon", false); if (temp == "not found") { temp = "0.000000000001"; } - convert(temp, epsilon); + m->mothurConvert(temp, epsilon); if (mindim < 1) { m->mothurOut("mindim must be at least 1."); m->mothurOutEndLine(); abort = true; } - if (maxdim < mindim) { m->mothurOut("maxdim must be greater than mindim."); m->mothurOutEndLine(); abort = true; } + if (maxdim < mindim) { maxdim = mindim; } } } @@ -153,32 +184,10 @@ NMDSCommand::NMDSCommand(string option) { } } //********************************************************************************************************************** -void NMDSCommand::help(){ - try { - m->mothurOut("The nmds command is modelled after the nmds code written in R by Sarah Goslee, using Non-metric multidimensional scaling function using the majorization algorithm from Borg & Groenen 1997, Modern Multidimensional Scaling."); m->mothurOutEndLine(); - m->mothurOut("The nmds command parameters are phylip, axes, mindim, maxdim, maxiters, iters and epsilon."); m->mothurOutEndLine(); - m->mothurOut("The phylip parameter allows you to enter your distance file."); m->mothurOutEndLine(); - m->mothurOut("The axes parameter allows you to enter a file containing a starting configuration."); m->mothurOutEndLine(); - m->mothurOut("The maxdim parameter allows you to select how maximum dimensions to use. Default=2"); m->mothurOutEndLine(); - m->mothurOut("The mindim parameter allows you to select how minimum dimensions to use. Default=2"); m->mothurOutEndLine(); - m->mothurOut("The maxiters parameter allows you to select the maximum number of iters to try with each random configuration. Default=500"); m->mothurOutEndLine(); - m->mothurOut("The iters parameter allows you to select the number of random configuration to try. Default=10"); m->mothurOutEndLine(); - m->mothurOut("The epsilon parameter allows you to select set an acceptable stopping point. Default=1e-12."); m->mothurOutEndLine(); - m->mothurOut("Example nmds(phylip=yourDistanceFile).\n"); - m->mothurOut("Note: No spaces between parameter labels (i.e. phylip), '=' and parameters (i.e.yourDistanceFile).\n\n"); - } - catch(exception& e) { - m->errorOut(e, "NMDSCommand", "help"); - exit(1); - } -} -//********************************************************************************************************************** -NMDSCommand::~NMDSCommand(){} -//********************************************************************************************************************** int NMDSCommand::execute(){ try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint); @@ -195,8 +204,10 @@ int NMDSCommand::execute(){ vector< vector > axes; if (axesfile != "") { axes = readAxes(names); } - string outputFileName = outputDir + m->getRootName(m->getSimpleName(phylipfile)) + "nmds.iters"; - string stressFileName = outputDir + m->getRootName(m->getSimpleName(phylipfile)) + "stress.nmds"; + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(phylipfile)); + string outputFileName = getOutputFileName("iters",variables); + string stressFileName = getOutputFileName("stress",variables); outputNames.push_back(outputFileName); outputTypes["iters"].push_back(outputFileName); outputNames.push_back(stressFileName); outputTypes["stress"].push_back(stressFileName); @@ -209,10 +220,12 @@ int NMDSCommand::execute(){ out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint); - out2 << "Dimension\tIter\tStress\tCorr" << endl; + out2 << "Dimension\tIter\tStress\tRsq" << endl; double bestStress = 10000000; + double bestR2 = 10000000; vector< vector > bestConfig; + int bestDim = 0; for (int i = mindim; i <= maxdim; i++) { m->mothurOut("Processing Dimension: " + toString(i)); m->mothurOutEndLine(); @@ -224,46 +237,52 @@ int NMDSCommand::execute(){ vector< vector > thisConfig; if (axesfile == "") { thisConfig = generateStartingConfiguration(names.size(), i); } else { thisConfig = getConfiguration(axes, i); } - if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { remove(outputNames[k].c_str()); } return 0; } + if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { m->mothurRemove(outputNames[k]); } return 0; } //calc nmds for this dimension double stress; vector< vector > endConfig = nmdsCalc(matrix, thisConfig, stress); - if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { remove(outputNames[k].c_str()); } return 0; } + if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { m->mothurRemove(outputNames[k]); } return 0; } //calc euclid distances for new config vector< vector > newEuclid = linearCalc.calculateEuclidianDistance(endConfig); - if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { remove(outputNames[k].c_str()); } return 0; } + if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { m->mothurRemove(outputNames[k]); } return 0; } //calc correlation between original distances and euclidean distances from this config - double corr = linearCalc.calcPearson(newEuclid, matrix); - corr *= corr; - if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { remove(outputNames[k].c_str()); } return 0; } + double rsquared = linearCalc.calcPearson(newEuclid, matrix); + rsquared *= rsquared; + if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { m->mothurRemove(outputNames[k]); } return 0; } //output results out << "Config" << (j+1) << '\t'; for (int k = 0; k < i; k++) { out << "axis" << (k+1) << '\t'; } out << endl; - out2 << i << '\t' << (j+1) << '\t' << stress << '\t' << corr << endl; + out2 << i << '\t' << (j+1) << '\t' << stress << '\t' << rsquared << endl; output(endConfig, names, out); //save best if (stress < bestStress) { + bestDim = i; bestStress = stress; + bestR2 = rsquared; bestConfig = endConfig; } - if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { remove(outputNames[k].c_str()); } return 0; } + if (m->control_pressed) { out.close(); out2.close(); for (int k = 0; k < outputNames.size(); k++) { m->mothurRemove(outputNames[k]); } return 0; } } } out.close(); out2.close(); //output best config - string BestFileName = outputDir + m->getRootName(m->getSimpleName(phylipfile)) + "nmds.axes"; + string BestFileName = getOutputFileName("nmds",variables); outputNames.push_back(BestFileName); outputTypes["nmds"].push_back(BestFileName); + m->mothurOut("\nNumber of dimensions:\t" + toString(bestDim) + "\n"); + m->mothurOut("Lowest stress :\t" + toString(bestStress) + "\n"); + m->mothurOut("R-squared for configuration:\t" + toString(bestR2) + "\n"); + ofstream outBest; m->openOutputFile(BestFileName, outBest); outBest.setf(ios::fixed, ios::floatfield); @@ -277,7 +296,7 @@ int NMDSCommand::execute(){ outBest.close(); - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine();