#include "ordervector.hpp"
#include "listvector.hpp"
+//sorts highest to lowest
+/***********************************************************************/
+inline bool abundNamesSort(string left, string right){
+
+ int countLeft = 0;
+ if(left != ""){
+ countLeft = 1;
+ for(int i=0;i<left.size();i++){ if(left[i] == ','){ countLeft++; } }
+ }
+
+ int countRight = 0;
+ if(right != ""){
+ countRight = 1;
+ for(int i=0;i<right.size();i++){ if(right[i] == ','){ countRight++; } }
+ }
+
+ if (countLeft > countRight) {
+ return true;
+ }
+ return false;
+}
/***********************************************************************/
try {
for(int i=0;i<data.size();i++){
if(data[i] != ""){
- int binSize = getNumNames(data[i]);
+ int binSize = m->getNumNames(data[i]);
numBins = i+1;
if(binSize > maxRank) { maxRank = binSize; }
numSeqs += binSize;
ListVector::ListVector(ifstream& f) : DataVector(), maxRank(0), numBins(0), numSeqs(0) {
try {
int hold;
- f >> label >> hold;
+
+ //are we at the beginning of the file??
+ if (m->saveNextLabel == "") {
+ f >> label;
+
+ //is this a shared file that has headers
+ if (label == "label") {
+
+ //gets "numOtus"
+ f >> label; m->gobble(f);
+
+ //eat rest of line
+ label = m->getline(f); m->gobble(f);
+
+ //parse labels to save
+ istringstream iStringStream(label);
+ m->listBinLabelsInFile.clear();
+ while(!iStringStream.eof()){
+ if (m->control_pressed) { break; }
+ string temp;
+ iStringStream >> temp; m->gobble(iStringStream);
+
+ m->listBinLabelsInFile.push_back(temp);
+ }
+
+ f >> label >> hold;
+ }else {
+ //read in first row
+ f >> hold;
+
+ //make binlabels because we don't have any
+ string snumBins = toString(hold);
+ m->listBinLabelsInFile.clear();
+ for (int i = 0; i < hold; i++) {
+ //if there is a bin label use it otherwise make one
+ string binLabel = "Otu";
+ string sbinNumber = toString(i+1);
+ if (sbinNumber.length() < snumBins.length()) {
+ int diff = snumBins.length() - sbinNumber.length();
+ for (int h = 0; h < diff; h++) { binLabel += "0"; }
+ }
+ binLabel += sbinNumber;
+ m->listBinLabelsInFile.push_back(binLabel);
+ }
+ }
+ m->saveNextLabel = label;
+ }else {
+ f >> label >> hold;
+ m->saveNextLabel = label;
+ }
+ binLabels.assign(m->listBinLabelsInFile.begin(), m->listBinLabelsInFile.begin()+hold);
+
data.assign(hold, "");
string inputData = "";
f >> inputData;
set(i, inputData);
}
-
- gobble(f);
+ m->gobble(f);
+
+ if (f.eof()) { m->saveNextLabel = ""; }
}
catch(exception& e) {
m->errorOut(e, "ListVector", "ListVector");
void ListVector::set(int binNumber, string seqNames){
try {
- int nNames_old = getNumNames(data[binNumber]);
+ int nNames_old = m->getNumNames(data[binNumber]);
data[binNumber] = seqNames;
- int nNames_new = getNumNames(seqNames);
+ int nNames_new = m->getNumNames(seqNames);
if(nNames_old == 0) { numBins++; }
if(nNames_new == 0) { numBins--; }
string ListVector::get(int index){
return data[index];
}
+/***********************************************************************/
+
+void ListVector::setLabels(vector<string> labels){
+ try {
+ binLabels = labels;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ListVector", "setLabels");
+ exit(1);
+ }
+}
+
+/***********************************************************************/
+//could potentially end up with duplicate binlabel names with code below.
+//we don't currently use them in a way that would do that.
+//if you had a listfile that had been subsampled and then added to it, dup names would be possible.
+vector<string> ListVector::getLabels(){
+ try {
+
+ string tagHeader = "Otu";
+ if (m->sharedHeaderMode == "tax") { tagHeader = "PhyloType"; }
+
+ if (binLabels.size() < data.size()) {
+ string snumBins = toString(numBins);
+
+ for (int i = 0; i < numBins; i++) {
+ string binLabel = tagHeader;
+
+ if (i < binLabels.size()) { //label exists, check leading zeros length
+ string sbinNumber = m->getSimpleLabel(binLabels[i]);
+ if (sbinNumber.length() < snumBins.length()) {
+ int diff = snumBins.length() - sbinNumber.length();
+ for (int h = 0; h < diff; h++) { binLabel += "0"; }
+ }
+ binLabel += sbinNumber;
+ binLabels[i] = binLabel;
+ }else{
+ string sbinNumber = toString(i+1);
+ if (sbinNumber.length() < snumBins.length()) {
+ int diff = snumBins.length() - sbinNumber.length();
+ for (int h = 0; h < diff; h++) { binLabel += "0"; }
+ }
+ binLabel += sbinNumber;
+ binLabels.push_back(binLabel);
+ }
+ }
+ }
+ return binLabels;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ListVector", "getLabels");
+ exit(1);
+ }
+}
/***********************************************************************/
void ListVector::push_back(string seqNames){
try {
data.push_back(seqNames);
- int nNames = getNumNames(seqNames);
+ int nNames = m->getNumNames(seqNames);
numBins++;
}
+/***********************************************************************/
+void ListVector::printHeaders(ostream& output){
+ try {
+ string snumBins = toString(numBins);
+ output << "label\tnumOtus\t";
+ if (m->sharedHeaderMode == "tax") {
+ for (int i = 0; i < numBins; i++) {
+
+ //if there is a bin label use it otherwise make one
+ string binLabel = "PhyloType";
+ string sbinNumber = toString(i+1);
+ if (sbinNumber.length() < snumBins.length()) {
+ int diff = snumBins.length() - sbinNumber.length();
+ for (int h = 0; h < diff; h++) { binLabel += "0"; }
+ }
+ binLabel += sbinNumber;
+ if (i < binLabels.size()) { binLabel = binLabels[i]; }
+
+ output << binLabel << '\t';
+ }
+ output << endl;
+ }else {
+ for (int i = 0; i < numBins; i++) {
+ //if there is a bin label use it otherwise make one
+ string binLabel = "Otu";
+ string sbinNumber = toString(i+1);
+ if (sbinNumber.length() < snumBins.length()) {
+ int diff = snumBins.length() - sbinNumber.length();
+ for (int h = 0; h < diff; h++) { binLabel += "0"; }
+ }
+ binLabel += sbinNumber;
+ if (i < binLabels.size()) { binLabel = binLabels[i]; }
+
+ output << binLabel << '\t';
+ }
+
+ output << endl;
+ }
+ m->printedListHeaders = true;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "ListVector", "printHeaders");
+ exit(1);
+ }
+}
+
/***********************************************************************/
void ListVector::print(ostream& output){
try {
output << label << '\t' << numBins << '\t';
- for(int i=0;i<data.size();i++){
- if(data[i] != ""){
- output << data[i] << '\t';
+ vector<string> hold = data;
+ sort(hold.begin(), hold.end(), abundNamesSort);
+
+ for(int i=0;i<hold.size();i++){
+ if(hold[i] != ""){
+ output << hold[i] << '\t';
}
}
output << endl;
RAbundVector rav;
for(int i=0;i<data.size();i++){
- int binSize = getNumNames(data[i]);
+ int binSize = m->getNumNames(data[i]);
rav.push_back(binSize);
}
SAbundVector sav(maxRank+1);
for(int i=0;i<data.size();i++){
- int binSize = getNumNames(data[i]);
+ int binSize = m->getNumNames(data[i]);
sav.set(binSize, sav.get(binSize) + 1);
}
sav.set(0, 0);
OrderVector ov;
for(int i=0;i<data.size();i++){
- int binSize = getNumNames(data[i]);
+ int binSize = m->getNumNames(data[i]);
for(int j=0;j<binSize;j++){
ov.push_back(i);
}
projects / mothur.git / blobdiff