void toMatrix(const SimpleMatrix& from, Matrix* to) {
const int nr = from.nrow();
const int nc = from.ncol();
to->Dimension(nr, nc);
// copy value
for (int i = 0; i < nr; ++i) {
for (int j = 0; j < nc; ++j) {
(*to)(i, j) = from[i][j];
}
}
// copy col labels
for (int i = 0; i < nc; ++i) {
(*to).SetColumnLabel(i, from.getColName()[i].c_str());
}
}
/**
* Load covariate from @param fn, using specified @param covNameToUse, for
* given
* @param includedSample
* covariate will be stored in @param covariate, and column names will be
* stored
* in @colNames
* if covariate file missed some samples, those sample names will be stored in
* @sampleToDrop
* NOTE: for missing values in a covariate, it will drop this covariate out of
* the following anaylysis
* @return number of samples have covariates.
* Example:
* includedSample = [A, B, C] and in covaraite file we have [B, C, C, D]
* then output covariate have 3 rows corresponding to [A, B, C]
* row C filled by the last C in covariate file
* sample D will be in sampleToDrop
*/
int _loadCovariate(const std::string& fn,
const std::vector<std::string>& includedSample,
const std::vector<std::string>& covNameToUse,
DataLoader::HandleMissingCov handleMissingCov,
SimpleMatrix* covariate, std::vector<std::string>* colNames,
std::set<std::string>* sampleToDrop) {
// load covariate
SimpleMatrix mat;
int ret = extractCovariate(fn, includedSample, covNameToUse, handleMissingCov,
&mat, sampleToDrop);
if (ret < 0) {
return -1;
}
// create covariate sample index
// const int nr = mat.nrow();
const int nc = mat.ncol();
std::map<std::string, int> covIndex;
makeMap(mat.getRowName(), &covIndex);
int idx = 0;
for (size_t i = 0; i < includedSample.size(); ++i) {
if (covIndex.find(includedSample[i]) == covIndex.end()) {
sampleToDrop->insert(includedSample[i]);
continue;
}
const int match = covIndex[includedSample[i]];
covariate->resize(idx + 1, nc);
for (int j = 0; j < mat.ncol(); ++j) {
(*covariate)[idx][j] = mat[match][j];
// skip row label, as MathMatrix class does not have row label
}
++idx;
}
// set col label
for (int i = 0; i < mat.ncol(); ++i) {
// (*covariate).SetColumnLabel(i, mat.getColName()[i].c_str());
(*covariate).setColName(i, mat.getColName()[i]);
}
return 0;
} // end _loadCovariate
