//' The reads in the read collection.
//'
//' This simply returns the full read count.
//'
//' @param acc An accession or a path to an actual SRA file (with .sra suffix)
//' @param n The number of reads to return
//' @return the reads in the collection
//' @export
//' @examples
//' reads('SRR000123')
// [[Rcpp::export]]
SEXP reads(Rcpp::String acc, int n, SEXP lkup) {
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
ReadIterator rgi = run.getReads( Read::all );
SEXP r_ans_width, r_ans;
XVectorList_holder r_ans_holder;
PROTECT(r_ans_width = IntegerVector(1,1000));
PROTECT(r_ans = alloc_XRawList("DNAStringSet", "DNAString", r_ans_width));
vector<std::string> out;
for(int i = 0; rgi.nextRead() & (i < 3) ; i++) {
cout << i;
while ( rgi.nextFragment() ) {
std::string str1 = rgi.getFragmentBases().toString();
const char * abc = str1.c_str();
cout << abc;
Chars_holder r_ans_elt_holder = get_elt_from_XRawList_holder(&r_ans_holder, i);
Ocopy_bytes_to_i1i2_with_lkup(0, r_ans_elt_holder.length - 1,
(char *)r_ans_elt_holder.ptr, r_ans_elt_holder.length,
abc, str1.length(),
INTEGER(lkup), LENGTH(lkup));
}
}
UNPROTECT(2);
return r_ans;
}
//' The readCount in the read collection.
//'
//' This simply returns the full read count.
//'
//' @param acc An accession or a path to an actual SRA file (with .sra suffix)
//' @return the number of reads in the collection
//' @export
//' @examples
//' getReference('SRR390728')
// [[Rcpp::export]]
DataFrame getReference(Rcpp::String acc) {
// open requested accession using SRA implementation of the API
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
Rcpp::String run_name ( run.getName () );
// get all references
ReferenceIterator it ( run.getReferences () );
vector<std::string> out;
vector<std::string> out2;
vector<long> out3;
vector<long> out4;
while( it.nextReference () ) {
out.push_back(it.getCanonicalName()) ;
out2.push_back(it.getCommonName()) ;
out3.push_back(it.getLength()) ;
out4.push_back(it.getAlignmentCount());
}
return DataFrame::create (
_["CanonicalNames"] = out, _["CommonNames"] = out2, _["Length"] = out3, _["AlignmentCount"] = out4
);
}
static void run ( String acc, String refname, int start, int stop )
{
// open requested accession using SRA implementation of the API
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
String run_name = run.getName ();
// get requested reference
Reference ref = run.getReference ( refname );
// start iterator on requested range
long count = stop - start + 1;
AlignmentIterator it = ref.getAlignmentSlice ( start, count, Alignment::primaryAlignment );
long i;
for ( i = 0; it . nextAlignment (); ++ i )
{
cout << it.getReadId ()
<< '\t' << it.getReferenceSpec ()
<< '\t' << it.getAlignmentPosition ()
<< '\t' << it.getLongCigar ( false ) // unclipped
<< '\t' << it.getAlignedFragmentBases ()
<< '\n';
}
cerr << "Read " << i << " alignments for " << run_name << '\n';
}
static void run ( String acc, int splitNum, int splitNo )
{
// open requested accession using SRA implementation of the API
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
String run_name = run.getName ();
// compute window to iterate through
long MAX_ROW = run.getReadCount ();
double chunk = ( double ) MAX_ROW / splitNum;
long first = ( long ) round ( chunk * ( splitNo-1 ) );
long next_first = ( long ) round ( chunk * ( splitNo ) );
if ( next_first > MAX_ROW )
next_first = MAX_ROW;
//start iterator on reads
long count = next_first - first;
ReadIterator it = run.getReadRange ( first+1, count, Read::all );
long i;
for ( i = 0; it.nextRead (); ++ i )
{
cout << it.getReadId();
//iterate through fragments
while ( it.nextFragment () )
cout << '\t' << it.getFragmentBases ();
cout << '\n';
}
cerr << "Read " << i << " spots for " << run_name << '\n';
}
static
void
run ( const DumpArgs & TheArgs )
{
if ( ! TheArgs . good () ) {
throw ErrorMsg ( "Invalid arguments" );
}
ngs :: String Acc ( TheArgs . accession () . c_str () );
ReadCollection RCol = ncbi :: NGS :: openReadCollection ( Acc );
int64_t minSpot = TheArgs . minSpotId ();
int64_t maxSpot = TheArgs . maxSpotId ();
if ( minSpot == 0 ) {
minSpot = 1;
}
if ( maxSpot == 0 ) {
maxSpot = RCol . getReadCount ();
}
if ( maxSpot < minSpot ) {
int64_t Id = minSpot;
minSpot = maxSpot;
maxSpot = Id;
}
ReadIterator Iterator = RCol.getReadRange (
minSpot,
maxSpot - minSpot + 1,
TheArgs . category ()
);
ngs :: String ReadCollectionName = RCol.getName ();
AFilters Filters ( TheArgs . accession () );
setupFilters ( Filters, TheArgs );
for ( int64_t llp = TheArgs . minSpotId () ; Iterator.nextRead (); llp ++ ) {
if ( Filters . checkIt ( Iterator ) ) {
if ( TheArgs . fastaDump () ) {
dumpFastA ( llp, ReadCollectionName, Iterator, TheArgs . fastaDumpWidth () );
}
else {
dumpFastQ ( llp, ReadCollectionName, Iterator );
}
}
}
kout.flush ();
std :: cerr << Filters . report ( TheArgs . legacyReport () );
} /* run () */
//' The readCount in the read collection.
//'
//' This simply returns the full read count.
//'
//' @param acc An accession or a path to an actual SRA file (with .sra suffix)
//' @return the number of reads in the collection
//' @export
//' @examples
//' readCount('SRR000123')
// [[Rcpp::export]]
List readCount2(Rcpp::String acc) {
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
long alignCount = run.getAlignmentCount();
long readCount = run.getReadCount();
ngs::String name = run.getName();
List AlignCount = Rcpp::List::create(Rcpp::Named("RefName") = name,
Rcpp::Named("ReadCount") = readCount,
Rcpp::Named("AlignCount")= alignCount);
return AlignCount;
}
// This is a simple example of exporting a C++ function to R. You can
// source this function into an R session using the Rcpp::sourceCpp
// function (or via the Source button on the editor toolbar). Learn
// more about Rcpp at:
//
// http://www.rcpp.org/
// http://adv-r.had.co.nz/Rcpp.html
// http://gallery.rcpp.org/
//' The readCount in the read collection.
//'
//' This simply returns the full read count.
//'
//' @param acc An accession or a path to an actual SRA file (with .sra suffix)
//' @return the number of reads in the collection
//' @export
//'
// [[Rcpp::export]]
List refBases(Rcpp::String acc) {
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
ReferenceIterator ri = run.getReferences() ;
vector<std::string> rlist;
while ( ri.nextReference() ) {
rlist.push_back(ri.getReferenceBases(0));
}
return Rcpp::List::create(Rcpp::Named("referencebases") = rlist);
}
std::vector<std::vector<std::string>> getRefs(std::string acc) {
ReadCollection run = ncbi::NGS::openReadCollection ( acc );
Rcpp::String run_name ( run.getName () );
ReferenceIterator it ( run.getReferences () );
vector<std::string> out;
vector<std::string> out2;
while( it.nextReference () ) {
out.push_back(it.getCanonicalName()) ;
out2.push_back(it.getCommonName()) ;
}
std::vector<std::vector<std::string>> v;
v.push_back(out);
v.push_back(out2);
return v;
}
