X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=subsample.cpp;h=c55accd618991279bb3a1a0ff6d29e0d0250a81e;hb=96dbe925073caefaed6e6db85659c144a806aeb1;hp=457b7b9d5f6a1273b57e898b3adec90dff5dd9f0;hpb=72e0be6b9c80009d4dbee24e8d690ad9514dc6fb;p=mothur.git diff --git a/subsample.cpp b/subsample.cpp index 457b7b9..c55accd 100644 --- a/subsample.cpp +++ b/subsample.cpp @@ -8,62 +8,54 @@ #include "subsample.h" //********************************************************************************************************************** -Tree* SubSample::getSample(Tree* T, TreeMap* tmap, TreeMap* newTmap, int size, map originalNameMap) { +Tree* SubSample::getSample(Tree* T, CountTable* ct, CountTable* newCt, int size) { try { Tree* newTree = NULL; - map > newGroups; - vector subsampledSeqs = getSample(tmap, size, newGroups); + //remove seqs not in sample from counttable + vector Groups = ct->getNamesOfGroups(); + newCt->copy(ct); + newCt->addGroup("doNotIncludeMe"); - //remove seqs not in sample from treemap - for (map >::iterator it = newGroups.begin(); it != newGroups.end(); it++) { - for (int i = 0; i < (it->second).size(); i++) { - newTmap->addSeq((it->second)[i], it->first); - } - } - - newTree = new Tree(newTmap); - newTree->getCopy(T, originalNameMap); - - return newTree; - } - catch(exception& e) { - m->errorOut(e, "SubSample", "getSample-Tree"); - exit(1); - } -} -/********************************************************************************************************************** -Tree* SubSample::getSample(Tree* T, TreeMap* tmap, map whole, int size) { - try { - Tree* newTree = NULL; - - vector subsampledSeqs = getSample(tmap, size); - map sampledNameMap = deconvolute(whole, subsampledSeqs); + map doNotIncludeTotals; + vector namesSeqs = ct->getNamesOfSeqs(); + for (int i = 0; i < namesSeqs.size(); i++) { doNotIncludeTotals[namesSeqs[i]] = 0; } + + for (int i = 0; i < Groups.size(); i++) { + if (m->inUsersGroups(Groups[i], m->getGroups())) { + if (m->control_pressed) { break; } - //remove seqs not in sample from treemap - for (int i = 0; i < tmap->namesOfSeqs.size(); i++) { - //is that name in the subsample? - int count = 0; - for (int j = 0; j < subsampledSeqs.size(); j++) { - if (tmap->namesOfSeqs[i] == subsampledSeqs[j]) { break; } //found it - count++; + int thisSize = ct->getGroupCount(Groups[i]); + + if (thisSize >= size) { + + vector names = ct->getNamesOfSeqs(Groups[i]); + vector random; + for (int j = 0; j < names.size(); j++) { + int num = ct->getGroupCount(names[j], Groups[i]); + for (int k = 0; k < num; k++) { random.push_back(j); } + } + random_shuffle(random.begin(), random.end()); + + vector sampleRandoms; sampleRandoms.resize(names.size(), 0); + for (int j = 0; j < size; j++) { sampleRandoms[random[j]]++; } + for (int j = 0; j < sampleRandoms.size(); j++) { + newCt->setAbund(names[j], Groups[i], sampleRandoms[j]); + } + sampleRandoms.clear(); sampleRandoms.resize(names.size(), 0); + for (int j = size; j < thisSize; j++) { sampleRandoms[random[j]]++; } + for (int j = 0; j < sampleRandoms.size(); j++) { doNotIncludeTotals[names[j]] += sampleRandoms[j]; } + }else { m->mothurOut("[ERROR]: You have selected a size that is larger than "+Groups[i]+" number of sequences.\n"); m->control_pressed = true; } } - if (m->control_pressed) { return newTree; } - - //if you didnt find it, remove it - if (count == subsampledSeqs.size()) { - tmap->removeSeq(tmap->namesOfSeqs[i]); - i--; //need this because removeSeq removes name from namesOfSeqs - } } - //create new tree - int numUniques = sampledNameMap.size(); - if (sampledNameMap.size() == 0) { numUniques = subsampledSeqs.size(); } + for (map::iterator it = doNotIncludeTotals.begin(); it != doNotIncludeTotals.end(); it++) { + newCt->setAbund(it->first, "doNotIncludeMe", it->second); + } - newTree = new Tree(numUniques, tmap); //numNodes, treemap - newTree->getSubTree(T, subsampledSeqs, sampledNameMap); + newTree = new Tree(newCt); + newTree->getCopy(T, true); return newTree; } @@ -71,7 +63,7 @@ Tree* SubSample::getSample(Tree* T, TreeMap* tmap, map whole, in m->errorOut(e, "SubSample", "getSample-Tree"); exit(1); } -}*/ +} //********************************************************************************************************************** //assumes whole maps dupName -> uniqueName map SubSample::deconvolute(map whole, vector& wanted) { @@ -112,100 +104,6 @@ map SubSample::deconvolute(map whole, vector SubSample::getSample(TreeMap* tMap, int size, map >& sample) { - try { - vector temp2; - sample["doNotIncludeMe"] = temp2; - - vector namesInSample; - - vector Groups = tMap->getNamesOfGroups(); - for (int i = 0; i < Groups.size(); i++) { - - if (m->inUsersGroups(Groups[i], m->getGroups())) { - if (m->control_pressed) { break; } - - vector thisGroup; thisGroup.push_back(Groups[i]); - vector thisGroupsSeqs = tMap->getNamesSeqs(thisGroup); - int thisSize = thisGroupsSeqs.size(); - vector temp; - sample[Groups[i]] = temp; - - if (thisSize >= size) { - - random_shuffle(thisGroupsSeqs.begin(), thisGroupsSeqs.end()); - - for (int j = 0; j < size; j++) { sample[Groups[i]].push_back(thisGroupsSeqs[j]); namesInSample.push_back(thisGroupsSeqs[j]); } - for (int j = size; j < thisSize; j++) { sample["doNotIncludeMe"].push_back(thisGroupsSeqs[j]); } - - }else { m->mothurOut("[ERROR]: You have selected a size that is larger than "+Groups[i]+" number of sequences.\n"); m->control_pressed = true; } - } - } - - return namesInSample; - } - catch(exception& e) { - m->errorOut(e, "SubSample", "getSample-TreeMap"); - exit(1); - } -} - -//********************************************************************************************************************** -vector SubSample::getSample(TreeMap* tMap, int size) { - try { - vector sample; - - vector Groups = tMap->getNamesOfGroups(); - for (int i = 0; i < Groups.size(); i++) { - - if (m->inUsersGroups(Groups[i], m->getGroups())) { - if (m->control_pressed) { break; } - - vector thisGroup; thisGroup.push_back(Groups[i]); - vector thisGroupsSeqs = tMap->getNamesSeqs(thisGroup); - int thisSize = thisGroupsSeqs.size(); - - if (thisSize >= size) { - - random_shuffle(thisGroupsSeqs.begin(), thisGroupsSeqs.end()); - - for (int j = 0; j < size; j++) { sample.push_back(thisGroupsSeqs[j]); } - }else { m->mothurOut("[ERROR]: You have selected a size that is larger than "+Groups[i]+" number of sequences.\n"); m->control_pressed = true; } - } - } - - return sample; - } - catch(exception& e) { - m->errorOut(e, "SubSample", "getSample-TreeMap"); - exit(1); - } -} -//********************************************************************************************************************** -vector SubSample::getSample(TreeMap* tMap, vector Groups) { - try { - vector sample; - - //vector Groups = tMap->getNamesOfGroups(); - for (int i = 0; i < Groups.size(); i++) { - - if (m->control_pressed) { break; } - - vector thisGroup; thisGroup.push_back(Groups[i]); - vector thisGroupsSeqs = tMap->getNamesSeqs(thisGroup); - int thisSize = thisGroupsSeqs.size(); - - for (int j = 0; j < thisSize; j++) { sample.push_back(thisGroupsSeqs[j]); } - } - - return sample; - } - catch(exception& e) { - m->errorOut(e, "SubSample", "getSample-TreeMap"); - exit(1); - } -} -//********************************************************************************************************************** vector SubSample::getSample(vector& thislookup, int size) { try { @@ -250,7 +148,7 @@ vector SubSample::getSample(vector& thislookup, int //subsampling may have created some otus with no sequences in them eliminateZeroOTUS(thislookup); - + if (m->control_pressed) { return m->currentBinLabels; } //save mothurOut's binLabels to restore for next label @@ -322,8 +220,51 @@ int SubSample::eliminateZeroOTUS(vector& thislookup) { exit(1); } } +//********************************************************************************************************************** +int SubSample::getSample(SAbundVector*& sabund, int size) { + try { + + OrderVector* order = new OrderVector(); + *order = sabund->getOrderVector(NULL); + + int numBins = order->getNumBins(); + int thisSize = order->getNumSeqs(); + + if (thisSize > size) { + random_shuffle(order->begin(), order->end()); + + RAbundVector* rabund = new RAbundVector(numBins); + rabund->setLabel(order->getLabel()); - + for (int j = 0; j < size; j++) { + + if (m->control_pressed) { delete order; delete rabund; return 0; } + + int bin = order->get(j); + + int abund = rabund->get(bin); + rabund->set(bin, (abund+1)); + } + + delete sabund; + sabund = new SAbundVector(); + *sabund = rabund->getSAbundVector(); + delete rabund; + + }else if (thisSize < size) { m->mothurOut("[ERROR]: The size you requested is larger than the number of sequences in the sabund vector. You requested " + toString(size) + " and you only have " + toString(thisSize) + " seqs in your sabund vector.\n"); m->control_pressed = true; } + + if (m->control_pressed) { return 0; } + + delete order; + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "SubSampleCommand", "getSample"); + exit(1); + } +} //**********************************************************************************************************************