//[[Rcpp::export]]
Rcpp::RawVector toBitVec(Rcpp::LogicalVector indx) {
unsigned nBit = indx.size();
unsigned nByte = ceil(float(nBit)/8);
Rcpp::RawVector bytes(nByte);//default are all 0s
bytes.attr("bitlen") = nBit;
unsigned byteIndex, bitIndex ;
for(unsigned i = 0 ; i < nBit; i++) {
byteIndex = i / 8;
bitIndex = i % 8;
if(indx(i) == 1)
bytes[byteIndex] = bytes[byteIndex] | 1 << bitIndex;
}
return bytes;
}
// [[Rcpp::export]]
Rcpp::DataFrame gt_to_popsum(Rcpp::DataFrame var_info, Rcpp::CharacterMatrix gt) {
// Calculate popgen summaries for the sample.
// var_info should contain columns named 'CHROM', 'POS', 'mask' and possibly others.
Rcpp::LogicalVector mask = var_info["mask"];
Rcpp::IntegerVector nsample(mask.size());
Rcpp::StringVector allele_counts(mask.size());
Rcpp::NumericVector Hes(mask.size());
Rcpp::NumericVector Nes(mask.size());
int i = 0;
int j = 0;
int k = 0;
for(i=0; i < gt.nrow(); i++){ // Iterate over variants (rows)
if(mask[i] == TRUE){
std::vector<int> myalleles (1,0);
for(j=0; j < gt.ncol(); j++){ // Iterate over samples (columns)
if(gt(i, j) != NA_STRING){
nsample[i]++; // Increment sample count.
// Count alleles per sample.
std::vector < int > intv = gtsplit(as<std::string>(gt(i, j)));
for(k=0; k<intv.size(); k++){
while(myalleles.size() - 1 < intv[k]){
// We have more alleles than exist in the vector myalleles.
myalleles.push_back(0);
}
myalleles[intv[k]]++;
}
}
}
// Concatenate allele counts into a comma delimited string.
int n;
char buffer [50];
n = sprintf (buffer, "%d", myalleles[0]);
for(j=1; j < myalleles.size(); j++){
n=sprintf (buffer, "%s,%d", buffer, myalleles[j]);
}
allele_counts[i] = buffer;
// Sum all alleles.
int nalleles = myalleles[0];
for(j=1; j < myalleles.size(); j++){
nalleles = nalleles + myalleles[j];
}
// Stats.
double He = 1;
He = He - pow(double(myalleles[0])/double(nalleles), myalleles.size());
for(j=1; j < myalleles.size(); j++){
He = He - pow(double(myalleles[j])/double(nalleles), myalleles.size());
}
Hes[i] = He;
Nes[i] = 1/(1-He);
} else { // Missing variant (row=NA)
nsample[i] = NA_INTEGER;
}
}
return Rcpp::DataFrame::create(var_info,
_["n"]=nsample,
_["Allele_counts"]=allele_counts,
_["He"]=Hes,
_["Ne"]=Nes
);
}
//' @export
// [[Rcpp::export(name=".gt_to_popsum")]]
Rcpp::DataFrame gt_to_popsum(Rcpp::DataFrame var_info, Rcpp::CharacterMatrix gt) {
// Calculate popgen summaries for the sample.
// var_info should contain columns named 'CHROM', 'POS', 'mask' and possibly others.
Rcpp::LogicalVector mask = var_info["mask"];
Rcpp::IntegerVector nsample(mask.size());
Rcpp::StringVector allele_counts(mask.size());
Rcpp::NumericVector Hes(mask.size());
Rcpp::NumericVector Nes(mask.size());
int i = 0;
int j = 0;
// unsigned int j = 0;
unsigned int k = 0;
for(i=0; i < gt.nrow(); i++){ // Iterate over variants (rows)
if(mask[i] == TRUE){
std::vector<int> myalleles (1,0);
for(j=0; j < gt.ncol(); j++){ // Iterate over samples (columns)
if(gt(i, j) != NA_STRING){
nsample[i]++; // Increment sample count.
// Rcout << "gt: " << gt(i, j) << "\n";
// Count alleles per sample.
int unphased_as_na = 0; // 0 == FALSE
std::vector < std::string > gt_vector;
std::string gt2 = as<std::string>(gt(i,j));
vcfRCommon::gtsplit( gt2, gt_vector, unphased_as_na );
// Rcout << "gt_vector.size: " << gt_vector.size() << "\n";
for(k=0; k<gt_vector.size(); k++){
int myAllele = std::stoi(gt_vector[k]);
// Rcout << " " << myAllele;
// // If this genotype had an allele we did not previously observe
// we'll have to grow the vector.
while(myalleles.size() - 1 < myAllele){
myalleles.push_back(0);
}
myalleles[myAllele]++;
}
// Rcout << "\n\n";
}
}
// Concatenate allele counts into a comma delimited string.
char buffer [50];
// int n;
// n=sprintf(buffer, "%d", myalleles[0]);
sprintf(buffer, "%d", myalleles[0]);
for(j=1; (unsigned)j < myalleles.size(); j++){
// n=sprintf (buffer, "%s,%d", buffer, myalleles[j]);
sprintf (buffer, "%s,%d", buffer, myalleles[j]);
}
allele_counts[i] = buffer;
// Sum all alleles.
int nalleles = myalleles[0];
for(j=1; (unsigned)j < myalleles.size(); j++){
nalleles = nalleles + myalleles[j];
}
// Stats.
double He = 1;
He = He - pow(double(myalleles[0])/double(nalleles), myalleles.size());
for(j=1; (unsigned)j < myalleles.size(); j++){
He = He - pow(double(myalleles[j])/double(nalleles), myalleles.size());
}
Hes[i] = He;
Nes[i] = 1/(1-He);
} else { // Missing variant (row=NA)
nsample[i] = NA_INTEGER;
}
}
return Rcpp::DataFrame::create(var_info,
_["n"]=nsample,
_["Allele_counts"]=allele_counts,
_["He"]=Hes,
_["Ne"]=Nes
);
}
