From d051154d778b77ba36dc603bbdb3c148a62a8e33 Mon Sep 17 00:00:00 2001 From: westcott Date: Fri, 17 Sep 2010 17:06:42 +0000 Subject: [PATCH] bug fixes --- metastats.h | 42 ++++++ metastatscommand.cpp | 323 +++++++++++++++++++++++++++++++++++++++++++ metastatscommand.h | 47 +++++++ 3 files changed, 412 insertions(+) create mode 100644 metastats.h create mode 100644 metastatscommand.cpp create mode 100644 metastatscommand.h diff --git a/metastats.h b/metastats.h new file mode 100644 index 0000000..ca8943a --- /dev/null +++ b/metastats.h @@ -0,0 +1,42 @@ +#ifndef METASTATS2 +#define METASTATS2 + +/* + * metastats.h + * Mothur + * + * Created by westcott on 9/16/10. + * Copyright 2010 Schloss Lab. All rights reserved. + * + */ + +#ifdef __cplusplus +extern "C" { +#endif + +#include +#include +#include +#include +#include +#include "fisher2.h" + +void testp(double *permuted_ttests,int *B,double *permuted,double + *Imatrix,int *nc,int *nr,int *g,double *Tinitial,double *ps); +void permute_matrix(double *Imatrix,int *nc,int *nr,double + *permuted,int *g,double *trial_ts,double *Tinitial,double + *counter); +void permute_array(int *array, int n); +void calc_twosample_ts(double *Pmatrix,int *g,int *nc,int *nr,double + *Ts,double *Tinitial,double *counter1); +void meanvar(double *pmatrix,int *g,int *nr,int *nc,double *storage); +void start(double *Imatrix,int *g,int *nr,int *nc,double *testing, + double storage[][9]); + +int metastat_main (char* outputFileName, int numRows, int numCols, double threshold, int numPermutations, double** data, int); + +#ifdef __cplusplus +} +#endif + +#endif \ No newline at end of file diff --git a/metastatscommand.cpp b/metastatscommand.cpp new file mode 100644 index 0000000..01e2606 --- /dev/null +++ b/metastatscommand.cpp @@ -0,0 +1,323 @@ +/* + * metastatscommand.cpp + * Mothur + * + * Created by westcott on 9/16/10. + * Copyright 2010 Schloss Lab. All rights reserved. + * + */ + +#include "metastatscommand.h" +#include "metastats.h" + +//********************************************************************************************************************** + +MetaStatsCommand::MetaStatsCommand(string option) { + try { + globaldata = GlobalData::getInstance(); + abort = false; + allLines = 1; + labels.clear(); + + //allow user to run help + if(option == "help") { help(); abort = true; } + + else { + //valid paramters for this command + string AlignArray[] = {"groups","label","outputdir","iters","threshold","g","design","inputdir"}; + vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); + + OptionParser parser(option); + map parameters = parser.getParameters(); + + ValidParameters validParameter; + + //check to make sure all parameters are valid for command + map::iterator it; + for (it = parameters.begin(); it != parameters.end(); it++) { + if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } + } + + //if the user changes the output directory command factory will send this info to us in the output parameter + outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ + outputDir = ""; + outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it + } + + //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 = ""; } + else { + string path; + it = parameters.find("design"); + //user has given a template file + if(it != parameters.end()){ + path = m->hasPath(it->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["design"] = inputDir + it->second; } + } + } + + //check for required parameters + designfile = validParameter.validFile(parameters, "design", true); + if (designfile == "not open") { abort = true; } + else if (designfile == "not found") { designfile = ""; m->mothurOut("You must provide an design file."); m->mothurOutEndLine(); abort = true; } + + //make sure the user has already run the read.otu command + if ((globaldata->getSharedFile() == "")) { + m->mothurOut("You must read a list and a group, or a shared file before you can use the normalize.shared command."); m->mothurOutEndLine(); abort = true; + } + + //check for optional parameter and set defaults + // ...at some point should added some additional type checking... + label = validParameter.validFile(parameters, "label", false); + if (label == "not found") { label = ""; } + else { + if(label != "all") { m->splitAtDash(label, labels); allLines = 0; } + else { allLines = 1; } + } + + //if the user has not specified any labels use the ones from read.otu + if (label == "") { + allLines = globaldata->allLines; + labels = globaldata->labels; + } + + groups = validParameter.validFile(parameters, "groups", false); + if (groups == "not found") { groups = ""; pickedGroups = false; } + else { + pickedGroups = true; + m->splitAtDash(groups, Groups); + globaldata->Groups = Groups; + } + + string temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; } + convert(temp, iters); + + temp = validParameter.validFile(parameters, "threshold", false); if (temp == "not found") { temp = "0.05"; } + convert(temp, threshold); + + //temp = validParameter.validFile(parameters, "g", false); if (temp == "not found") { temp = "0"; } + //convert(temp, g); + } + + } + catch(exception& e) { + m->errorOut(e, "MetaStatsCommand", "MetaStatsCommand"); + exit(1); + } +} + +//********************************************************************************************************************** + +void MetaStatsCommand::help(){ + try { + m->mothurOut("Paper reference goes here....\n"); + m->mothurOut("The metastats command can only be executed after a successful read.otu command of a list and group or shared file.\n"); + m->mothurOut("The metastats command outputs a .... file. \n"); + m->mothurOut("The metastats command parameters are iters, threshold, groups, design and label. No parameters are required.\n"); + m->mothurOut("The design parameter allows you to assign your groups to sets when you are running metastat. mothur will run all pairwise comparisons of the sets. It is required. \n"); + m->mothurOut("The design file looks like the group file. It is a 2 column tab delimited file, where the first column is the group name and the second column is the set the group belongs to.\n"); + m->mothurOut("The iters parameter ...\n"); + m->mothurOut("The threshold parameter ...\n"); + m->mothurOut("The groups parameter allows you to specify which of the groups in your groupfile you would like included. The group names are separated by dashes.\n"); + m->mothurOut("The label parameter allows you to select what distance levels you would like, and are also separated by dashes.\n"); + m->mothurOut("The metastats command should be in the following format: metastats(groups=yourGroups, label=yourLabels).\n"); + m->mothurOut("Example metastats(groups=A-B-C).\n"); + m->mothurOut("The default value for groups is all the groups in your groupfile, and all labels in your inputfile will be used.\n"); + m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n"); + + } + catch(exception& e) { + m->errorOut(e, "MetaStatsCommand", "help"); + exit(1); + } +} + +//********************************************************************************************************************** + +MetaStatsCommand::~MetaStatsCommand(){} + +//********************************************************************************************************************** + +int MetaStatsCommand::execute(){ + try { + + if (abort == true) { return 0; } + + designMap = new GroupMap(designfile); + designMap->readDesignMap(); + + read = new ReadOTUFile(globaldata->inputFileName); + read->read(&*globaldata); + input = globaldata->ginput; + lookup = input->getSharedRAbundVectors(); + string lastLabel = lookup[0]->getLabel(); + + //if the users enters label "0.06" and there is no "0.06" in their file use the next lowest label. + set processedLabels; + set userLabels = labels; + + //as long as you are not at the end of the file or done wih the lines you want + while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) { + + if (m->control_pressed) { for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } globaldata->Groups.clear(); delete read; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + + if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){ + + m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine(); + process(lookup); + + processedLabels.insert(lookup[0]->getLabel()); + userLabels.erase(lookup[0]->getLabel()); + } + + if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) { + string saveLabel = lookup[0]->getLabel(); + + for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } + lookup = input->getSharedRAbundVectors(lastLabel); + m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine(); + + process(lookup); + + processedLabels.insert(lookup[0]->getLabel()); + userLabels.erase(lookup[0]->getLabel()); + + //restore real lastlabel to save below + lookup[0]->setLabel(saveLabel); + } + + lastLabel = lookup[0]->getLabel(); + //prevent memory leak + for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; lookup[i] = NULL; } + + if (m->control_pressed) { globaldata->Groups.clear(); delete read; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + + //get next line to process + lookup = input->getSharedRAbundVectors(); + } + + if (m->control_pressed) { globaldata->Groups.clear(); delete read; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + + //output error messages about any remaining user labels + set::iterator it; + bool needToRun = false; + for (it = userLabels.begin(); it != userLabels.end(); it++) { + m->mothurOut("Your file does not include the label " + *it); + if (processedLabels.count(lastLabel) != 1) { + m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine(); + needToRun = true; + }else { + m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine(); + } + } + + //run last label if you need to + if (needToRun == true) { + for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } } + lookup = input->getSharedRAbundVectors(lastLabel); + + m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine(); + + process(lookup); + + for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } + } + + //reset groups parameter + globaldata->Groups.clear(); + delete input; globaldata->ginput = NULL; + delete read; + + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0;} + + m->mothurOutEndLine(); + m->mothurOut("Output File Names: "); m->mothurOutEndLine(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } + m->mothurOutEndLine(); + + return 0; + } + catch(exception& e) { + m->errorOut(e, "MetaStatsCommand", "execute"); + exit(1); + } +} +//********************************************************************************************************************** + +int MetaStatsCommand::process(vector& thisLookUp){ + try { + if (pickedGroups) { eliminateZeroOTUS(thisLookUp); } + + string outputFileName = outputDir + m->getRootName(m->getSimpleName(globaldata->inputFileName)) + thisLookUp[0]->getLabel() + ".metastats"; + outputNames.push_back(outputFileName); + int nameLength = outputFileName.length(); + char * output = new char[nameLength]; + strcpy(output, outputFileName.c_str()); + + //build matrix from shared rabunds + double** data; + data = new double*[thisLookUp[0]->getNumBins()]; + + for (int j = 0; j < thisLookUp[0]->getNumBins(); j++) { + data[j] = new double[thisLookUp.size()]; + for (int i = 0; i < thisLookUp.size(); i++) { + data[j][i] = (thisLookUp[i]->getAbundance(j)); + } + } + + if (g == 0) { g = thisLookUp.size() / 2; } + + metastat_main(output, thisLookUp[0]->getNumBins(), thisLookUp.size(), threshold, iters, data, g); + + return 0; + } + catch(exception& e) { + m->errorOut(e, "MetaStatsCommand", "normalize"); + exit(1); + } +} +//********************************************************************************************************************** +int MetaStatsCommand::eliminateZeroOTUS(vector& thislookup) { + try { + + vector newLookup; + for (int i = 0; i < thislookup.size(); i++) { + SharedRAbundVector* temp = new SharedRAbundVector(); + temp->setLabel(thislookup[i]->getLabel()); + temp->setGroup(thislookup[i]->getGroup()); + newLookup.push_back(temp); + } + + //for each bin + for (int i = 0; i < thislookup[0]->getNumBins(); i++) { + if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; } + + //look at each sharedRabund and make sure they are not all zero + bool allZero = true; + for (int j = 0; j < thislookup.size(); j++) { + if (thislookup[j]->getAbundance(i) != 0) { allZero = false; break; } + } + + //if they are not all zero add this bin + if (!allZero) { + for (int j = 0; j < thislookup.size(); j++) { + newLookup[j]->push_back(thislookup[j]->getAbundance(i), thislookup[j]->getGroup()); + } + } + } + + for (int j = 0; j < thislookup.size(); j++) { delete thislookup[j]; } + + thislookup = newLookup; + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "MetaStatsCommand", "eliminateZeroOTUS"); + exit(1); + } +} +//********************************************************************************************************************** diff --git a/metastatscommand.h b/metastatscommand.h new file mode 100644 index 0000000..9c9f455 --- /dev/null +++ b/metastatscommand.h @@ -0,0 +1,47 @@ +#ifndef METASTATSCOMMAND_H +#define METASTATSCOMMAND_H + +/* + * metastatscommand.h + * Mothur + * + * Created by westcott on 9/16/10. + * Copyright 2010 Schloss Lab. All rights reserved. + * + */ + +#include "command.hpp" +#include "inputdata.h" +#include "readotu.h" +#include "sharedrabundvector.h" + +class GlobalData; + +class MetaStatsCommand : public Command { + +public: + MetaStatsCommand(string); + ~MetaStatsCommand(); + int execute(); + void help(); + +private: + GlobalData* globaldata; + GroupMap* designMap; + ReadOTUFile* read; + InputData* input; + vector lookup; + + bool abort, allLines, pickedGroups; + set labels; //holds labels to be used + string groups, label, outputDir, inputDir, designfile; + vector Groups, outputNames; + int iters, g; + float threshold; + + int process(vector&); + int eliminateZeroOTUS(vector&); +}; + +#endif + -- 2.39.2