#include "shannon.h"
#include "jackknife.h"
#include "coverage.h"
+#include "shannonrange.h"
//**********************************************************************************************************************
vector<string> RareFactCommand::setParameters(){
try {
- CommandParameter plist("list", "InputTypes", "", "", "LRSS", "LRSS", "none",false,false); parameters.push_back(plist);
- CommandParameter prabund("rabund", "InputTypes", "", "", "LRSS", "LRSS", "none",false,false); parameters.push_back(prabund);
- CommandParameter psabund("sabund", "InputTypes", "", "", "LRSS", "LRSS", "none",false,false); parameters.push_back(psabund);
- CommandParameter pshared("shared", "InputTypes", "", "", "LRSS", "LRSS", "none",false,false); parameters.push_back(pshared);
- CommandParameter plabel("label", "String", "", "", "", "", "",false,false); parameters.push_back(plabel);
- CommandParameter pfreq("freq", "Number", "", "100", "", "", "",false,false); parameters.push_back(pfreq);
- CommandParameter piters("iters", "Number", "", "1000", "", "", "",false,false); parameters.push_back(piters);
- CommandParameter pcalc("calc", "Multiple", "sobs-chao-nseqs-coverage-ace-jack-shannon-shannoneven-npshannon-heip-smithwilson-simpson-simpsoneven-invsimpson-bootstrap", "sobs", "", "", "",true,false); parameters.push_back(pcalc);
- CommandParameter pabund("abund", "Number", "", "10", "", "", "",false,false); parameters.push_back(pabund);
- CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
- CommandParameter pgroupmode("groupmode", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pgroupmode);
- CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
- CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
+ CommandParameter plist("list", "InputTypes", "", "", "LRSS", "LRSS", "none","",false,false,true); parameters.push_back(plist);
+ CommandParameter prabund("rabund", "InputTypes", "", "", "LRSS", "LRSS", "none","",false,false); parameters.push_back(prabund);
+ CommandParameter psabund("sabund", "InputTypes", "", "", "LRSS", "LRSS", "none","",false,false); parameters.push_back(psabund);
+ CommandParameter pshared("shared", "InputTypes", "", "", "LRSS", "LRSS", "none","",false,false,true); parameters.push_back(pshared);
+ CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
+ CommandParameter pfreq("freq", "Number", "", "100", "", "", "","",false,false); parameters.push_back(pfreq);
+ CommandParameter piters("iters", "Number", "", "1000", "", "", "","",false,false); parameters.push_back(piters);
+ CommandParameter pcalc("calc", "Multiple", "sobs-chao-nseqs-coverage-ace-jack-shannon-shannoneven-npshannon-heip-smithwilson-simpson-simpsoneven-invsimpson-bootstrap-shannonrange", "sobs", "", "", "","",true,false,true); parameters.push_back(pcalc);
+ CommandParameter pabund("abund", "Number", "", "10", "", "", "","",false,false); parameters.push_back(pabund);
+ CommandParameter palpha("alpha", "Multiple", "0-1-2", "1", "", "", "","",false,false,true); parameters.push_back(palpha);
+ CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors);
+ CommandParameter pgroupmode("groupmode", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pgroupmode);
+ 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); }
try {
ValidCalculators validCalculator;
string helpString = "";
- helpString += "The rarefaction.single command parameters are list, sabund, rabund, shared, label, iters, freq, calc, processors and abund. list, sabund, rabund or shared is required unless you have a valid current file. \n";
+ helpString += "The rarefaction.single command parameters are list, sabund, rabund, shared, label, iters, freq, calc, processors, groupmode and abund. list, sabund, rabund or shared is required unless you have a valid current file. \n";
helpString += "The freq parameter is used indicate when to output your data, by default it is set to 100. But you can set it to a percentage of the number of sequence. For example freq=0.10, means 10%. \n";
helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n";
helpString += "The rarefaction.single command should be in the following format: \n";
helpString += "rarefaction.single(label=yourLabel, iters=yourIters, freq=yourFreq, calc=yourEstimators).\n";
helpString += "Example rarefaction.single(label=unique-.01-.03, iters=10000, freq=10, calc=sobs-rchao-race-rjack-rbootstrap-rshannon-rnpshannon-rsimpson).\n";
helpString += "The default values for iters is 1000, freq is 100, and calc is rarefaction which calculates the rarefaction curve for the observed richness.\n";
+ helpString += "The alpha parameter is used to set the alpha value for the shannonrange calculator.\n";
validCalculator.printCalc("rarefaction");
helpString += "If you are running rarefaction.single with a shared file and would like your results collated in one file, set groupmode=t. (Default=true).\n";
helpString += "The label parameter is used to analyze specific labels in your input.\n";
exit(1);
}
}
-
+//**********************************************************************************************************************
+string RareFactCommand::getOutputPattern(string type) {
+ try {
+ string pattern = "";
+ if (type == "rarefaction") { pattern = "[filename],rarefaction"; }
+ else if (type == "r_chao") { pattern = "[filename],r_chao"; }
+ else if (type == "r_ace") { pattern = "[filename],r_ace"; }
+ else if (type == "r_jack") { pattern = "[filename],r_jack"; }
+ else if (type == "r_shannon") { pattern = "[filename],r_shannon"; }
+ else if (type == "r_shannoneven") { pattern = "[filename],r_shannoneven"; }
+ else if (type == "r_smithwilson") { pattern = "[filename],r_smithwilson"; }
+ else if (type == "r_npshannon") { pattern = "[filename],r_npshannon"; }
+ else if (type == "r_shannonrange"){ pattern = "[filename],r_shannonrange"; }
+ else if (type == "r_simpson") { pattern = "[filename],r_simpson"; }
+ else if (type == "r_simpsoneven") { pattern = "[filename],r_simpsoneven"; }
+ else if (type == "r_invsimpson") { pattern = "[filename],r_invsimpson"; }
+ else if (type == "r_bootstrap") { pattern = "[filename],r_bootstrap"; }
+ else if (type == "r_coverage") { pattern = "[filename],r_coverage"; }
+ else if (type == "r_nseqs") { pattern = "[filename],r_nseqs"; }
+ else if (type == "r_heip") { pattern = "[filename],r_heip"; }
+ else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
+
+ return pattern;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "RareFactCommand", "getOutputPattern");
+ exit(1);
+ }
+}
//**********************************************************************************************************************
RareFactCommand::RareFactCommand(){
try {
outputTypes["r_jack"] = tempOutNames;
outputTypes["r_shannon"] = tempOutNames;
outputTypes["r_shannoneven"] = tempOutNames;
+ outputTypes["r_shannonrange"] = tempOutNames;
outputTypes["r_heip"] = tempOutNames;
outputTypes["r_smithwilson"] = tempOutNames;
outputTypes["r_npshannon"] = tempOutNames;
outputTypes["r_jack"] = tempOutNames;
outputTypes["r_shannon"] = tempOutNames;
outputTypes["r_shannoneven"] = tempOutNames;
+ outputTypes["r_shannonrange"] = tempOutNames;
outputTypes["r_heip"] = tempOutNames;
outputTypes["r_smithwilson"] = tempOutNames;
outputTypes["r_npshannon"] = tempOutNames;
temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
m->setProcessors(temp);
m->mothurConvert(temp, processors);
+
+ temp = validParameter.validFile(parameters, "alpha", false); if (temp == "not found") { temp = "1"; }
+ m->mothurConvert(temp, alpha);
+
+ if ((alpha != 0) && (alpha != 1) && (alpha != 2)) { m->mothurOut("[ERROR]: Not a valid alpha value. Valid values are 0, 1 and 2."); m->mothurOutEndLine(); abort=true; }
temp = validParameter.validFile(parameters, "groupmode", false); if (temp == "not found") { temp = "T"; }
groupMode = m->isTrue(temp);
if (abort == true) { if (calledHelp) { return 0; } return 2; }
+ map<string, set<int> > labelToEnds;
if ((format != "sharedfile")) { inputFileNames.push_back(inputfile); }
- else { inputFileNames = parseSharedFile(sharedfile); format = "rabund"; }
-
- if (m->control_pressed) { return 0; }
+ else { inputFileNames = parseSharedFile(sharedfile, labelToEnds); format = "rabund"; }
+
+ if (m->control_pressed) { return 0; }
map<int, string> file2Group; //index in outputNames[i] -> group
for (int p = 0; p < inputFileNames.size(); p++) {
int i;
ValidCalculators validCalculator;
+ map<string, string> variables;
+ variables["[filename]"] = fileNameRoot;
for (i=0; i<Estimators.size(); i++) {
if (validCalculator.isValidCalculator("rarefaction", Estimators[i]) == true) {
if (Estimators[i] == "sobs") {
- rDisplays.push_back(new RareDisplay(new Sobs(), new ThreeColumnFile(fileNameRoot+"rarefaction")));
- outputNames.push_back(fileNameRoot+"rarefaction"); outputTypes["rarefaction"].push_back(fileNameRoot+"rarefaction");
+ rDisplays.push_back(new RareDisplay(new Sobs(), new ThreeColumnFile(getOutputFileName("rarefaction",variables))));
+ outputNames.push_back(getOutputFileName("rarefaction",variables)); outputTypes["rarefaction"].push_back(getOutputFileName("rarefaction",variables));
}else if (Estimators[i] == "chao") {
- rDisplays.push_back(new RareDisplay(new Chao1(), new ThreeColumnFile(fileNameRoot+"r_chao")));
- outputNames.push_back(fileNameRoot+"r_chao"); outputTypes["r_chao"].push_back(fileNameRoot+"r_chao");
+ rDisplays.push_back(new RareDisplay(new Chao1(), new ThreeColumnFile(getOutputFileName("r_chao",variables))));
+ outputNames.push_back(getOutputFileName("r_chao",variables)); outputTypes["r_chao"].push_back(getOutputFileName("r_chao",variables));
}else if (Estimators[i] == "ace") {
if(abund < 5)
abund = 10;
- rDisplays.push_back(new RareDisplay(new Ace(abund), new ThreeColumnFile(fileNameRoot+"r_ace")));
- outputNames.push_back(fileNameRoot+"r_ace"); outputTypes["r_ace"].push_back(fileNameRoot+"r_ace");
+ rDisplays.push_back(new RareDisplay(new Ace(abund), new ThreeColumnFile(getOutputFileName("r_ace",variables))));
+ outputNames.push_back(getOutputFileName("r_ace",variables)); outputTypes["r_ace"].push_back(getOutputFileName("r_ace",variables));
}else if (Estimators[i] == "jack") {
- rDisplays.push_back(new RareDisplay(new Jackknife(), new ThreeColumnFile(fileNameRoot+"r_jack")));
- outputNames.push_back(fileNameRoot+"r_jack"); outputTypes["r_jack"].push_back(fileNameRoot+"r_jack");
+ rDisplays.push_back(new RareDisplay(new Jackknife(), new ThreeColumnFile(getOutputFileName("r_jack",variables))));
+ outputNames.push_back(getOutputFileName("r_jack",variables)); outputTypes["r_jack"].push_back(getOutputFileName("r_jack",variables));
}else if (Estimators[i] == "shannon") {
- rDisplays.push_back(new RareDisplay(new Shannon(), new ThreeColumnFile(fileNameRoot+"r_shannon")));
- outputNames.push_back(fileNameRoot+"r_shannon"); outputTypes["r_shannon"].push_back(fileNameRoot+"r_shannon");
+ rDisplays.push_back(new RareDisplay(new Shannon(), new ThreeColumnFile(getOutputFileName("r_shannon",variables))));
+ outputNames.push_back(getOutputFileName("r_shannon",variables)); outputTypes["r_shannon"].push_back(getOutputFileName("r_shannon",variables));
}else if (Estimators[i] == "shannoneven") {
- rDisplays.push_back(new RareDisplay(new ShannonEven(), new ThreeColumnFile(fileNameRoot+"r_shannoneven")));
- outputNames.push_back(fileNameRoot+"r_shannoneven"); outputTypes["r_shannoneven"].push_back(fileNameRoot+"r_shannoneven");
+ rDisplays.push_back(new RareDisplay(new ShannonEven(), new ThreeColumnFile(getOutputFileName("r_shannoneven",variables))));
+ outputNames.push_back(getOutputFileName("r_shannoneven",variables)); outputTypes["r_shannoneven"].push_back(getOutputFileName("r_shannoneven",variables));
}else if (Estimators[i] == "heip") {
- rDisplays.push_back(new RareDisplay(new Heip(), new ThreeColumnFile(fileNameRoot+"r_heip")));
- outputNames.push_back(fileNameRoot+"r_heip"); outputTypes["r_heip"].push_back(fileNameRoot+"r_heip");
- }else if (Estimators[i] == "smithwilson") {
- rDisplays.push_back(new RareDisplay(new SmithWilson(), new ThreeColumnFile(fileNameRoot+"r_smithwilson")));
- outputNames.push_back(fileNameRoot+"r_smithwilson"); outputTypes["r_smithwilson"].push_back(fileNameRoot+"r_smithwilson");
+ rDisplays.push_back(new RareDisplay(new Heip(), new ThreeColumnFile(getOutputFileName("r_heip",variables))));
+ outputNames.push_back(getOutputFileName("r_heip",variables)); outputTypes["r_heip"].push_back(getOutputFileName("r_heip",variables));
+ }else if (Estimators[i] == "r_shannonrange") {
+ rDisplays.push_back(new RareDisplay(new RangeShannon(alpha), new ThreeColumnFile(getOutputFileName("r_shannonrange", variables))));
+ outputNames.push_back(getOutputFileName("r_shannonrange", variables)); outputTypes["r_shannoneven"].push_back(getOutputFileName("r_shannonrange", variables));
+ }else if (Estimators[i] == "smithwilson") {
+ rDisplays.push_back(new RareDisplay(new SmithWilson(), new ThreeColumnFile(getOutputFileName("r_smithwilson",variables))));
+ outputNames.push_back(getOutputFileName("r_smithwilson",variables)); outputTypes["r_smithwilson"].push_back(getOutputFileName("r_smithwilson",variables));
}else if (Estimators[i] == "npshannon") {
- rDisplays.push_back(new RareDisplay(new NPShannon(), new ThreeColumnFile(fileNameRoot+"r_npshannon")));
- outputNames.push_back(fileNameRoot+"r_npshannon"); outputTypes["r_npshannon"].push_back(fileNameRoot+"r_npshannon");
+ rDisplays.push_back(new RareDisplay(new NPShannon(), new ThreeColumnFile(getOutputFileName("r_npshannon",variables))));
+ outputNames.push_back(getOutputFileName("r_npshannon",variables)); outputTypes["r_npshannon"].push_back(getOutputFileName("r_npshannon",variables));
}else if (Estimators[i] == "simpson") {
- rDisplays.push_back(new RareDisplay(new Simpson(), new ThreeColumnFile(fileNameRoot+"r_simpson")));
- outputNames.push_back(fileNameRoot+"r_simpson"); outputTypes["r_simpson"].push_back(fileNameRoot+"r_simpson");
+ rDisplays.push_back(new RareDisplay(new Simpson(), new ThreeColumnFile(getOutputFileName("r_simpson",variables))));
+ outputNames.push_back(getOutputFileName("r_simpson",variables)); outputTypes["r_simpson"].push_back(getOutputFileName("r_simpson",variables));
}else if (Estimators[i] == "simpsoneven") {
- rDisplays.push_back(new RareDisplay(new SimpsonEven(), new ThreeColumnFile(fileNameRoot+"r_simpsoneven")));
- outputNames.push_back(fileNameRoot+"r_simpsoneven"); outputTypes["r_simpsoneven"].push_back(fileNameRoot+"r_simpsoneven");
+ rDisplays.push_back(new RareDisplay(new SimpsonEven(), new ThreeColumnFile(getOutputFileName("r_simpsoneven",variables))));
+ outputNames.push_back(getOutputFileName("r_simpsoneven",variables)); outputTypes["r_simpsoneven"].push_back(getOutputFileName("r_simpsoneven",variables));
}else if (Estimators[i] == "invsimpson") {
- rDisplays.push_back(new RareDisplay(new InvSimpson(), new ThreeColumnFile(fileNameRoot+"r_invsimpson")));
- outputNames.push_back(fileNameRoot+"r_invsimpson"); outputTypes["r_invsimpson"].push_back(fileNameRoot+"r_invsimpson");
+ rDisplays.push_back(new RareDisplay(new InvSimpson(), new ThreeColumnFile(getOutputFileName("r_invsimpson",variables))));
+ outputNames.push_back(getOutputFileName("r_invsimpson",variables)); outputTypes["r_invsimpson"].push_back(getOutputFileName("r_invsimpson",variables));
}else if (Estimators[i] == "bootstrap") {
- rDisplays.push_back(new RareDisplay(new Bootstrap(), new ThreeColumnFile(fileNameRoot+"r_bootstrap")));
- outputNames.push_back(fileNameRoot+"r_bootstrap"); outputTypes["r_bootstrap"].push_back(fileNameRoot+"r_bootstrap");
+ rDisplays.push_back(new RareDisplay(new Bootstrap(), new ThreeColumnFile(getOutputFileName("r_bootstrap",variables))));
+ outputNames.push_back(getOutputFileName("r_bootstrap",variables)); outputTypes["r_bootstrap"].push_back(getOutputFileName("r_bootstrap",variables));
}else if (Estimators[i] == "coverage") {
- rDisplays.push_back(new RareDisplay(new Coverage(), new ThreeColumnFile(fileNameRoot+"r_coverage")));
- outputNames.push_back(fileNameRoot+"r_coverage"); outputTypes["r_coverage"].push_back(fileNameRoot+"r_coverage");
+ rDisplays.push_back(new RareDisplay(new Coverage(), new ThreeColumnFile(getOutputFileName("r_coverage",variables))));
+ outputNames.push_back(getOutputFileName("r_coverage",variables)); outputTypes["r_coverage"].push_back(getOutputFileName("r_coverage",variables));
}else if (Estimators[i] == "nseqs") {
- rDisplays.push_back(new RareDisplay(new NSeqs(), new ThreeColumnFile(fileNameRoot+"r_nseqs")));
- outputNames.push_back(fileNameRoot+"r_nseqs"); outputTypes["r_nseqs"].push_back(fileNameRoot+"r_nseqs");
+ rDisplays.push_back(new RareDisplay(new NSeqs(), new ThreeColumnFile(getOutputFileName("r_nseqs",variables))));
+ outputNames.push_back(getOutputFileName("r_nseqs",variables)); outputTypes["r_nseqs"].push_back(getOutputFileName("r_nseqs",variables));
}
if (inputFileNames.size() > 1) { file2Group[outputNames.size()-1] = groups[p]; }
}
if(allLines == 1 || labels.count(order->getLabel()) == 1){
m->mothurOut(order->getLabel()); m->mothurOutEndLine();
- rCurve = new Rarefact(order, rDisplays, processors);
+ map<string, set<int> >::iterator itEndings = labelToEnds.find(order->getLabel());
+ set<int> ends;
+ if (itEndings != labelToEnds.end()) { ends = itEndings->second; }
+ rCurve = new Rarefact(order, rDisplays, processors, ends);
rCurve->getCurve(freq, nIters);
delete rCurve;
order = (input->getOrderVector(lastLabel));
m->mothurOut(order->getLabel()); m->mothurOutEndLine();
- rCurve = new Rarefact(order, rDisplays, processors);
+ map<string, set<int> >::iterator itEndings = labelToEnds.find(order->getLabel());
+ set<int> ends;
+ if (itEndings != labelToEnds.end()) { ends = itEndings->second; }
+ rCurve = new Rarefact(order, rDisplays, processors, ends);
+
rCurve->getCurve(freq, nIters);
delete rCurve;
order = (input->getOrderVector(lastLabel));
m->mothurOut(order->getLabel()); m->mothurOutEndLine();
- rCurve = new Rarefact(order, rDisplays, processors);
+ map<string, set<int> >::iterator itEndings = labelToEnds.find(order->getLabel());
+ set<int> ends;
+ if (itEndings != labelToEnds.end()) { ends = itEndings->second; }
+ rCurve = new Rarefact(order, rDisplays, processors, ends);
+
rCurve->getCurve(freq, nIters);
delete rCurve;
}
}
//**********************************************************************************************************************
-vector<string> RareFactCommand::parseSharedFile(string filename) {
+vector<string> RareFactCommand::parseSharedFile(string filename, map<string, set<int> >& label2Ends) {
try {
vector<string> filenames;
m->openOutputFileAppend(sharedFileRoot + lookup[i]->getGroup() + ".rabund", *(filehandles[lookup[i]->getGroup()]));
rav.print(*(filehandles[lookup[i]->getGroup()]));
(*(filehandles[lookup[i]->getGroup()])).close();
+ label2Ends[lookup[i]->getLabel()].insert(rav.getNumSeqs());
}
for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
projects / mothur.git / blobdiff