// Splits a sam file into individual files, one per chromosome. The files are created in the specified directory.
// Each splitted file contains the same header as the original samFile. The generated files may only contain
// a header if no alignments to that chromosome exist. The names of the split files come from the sam header
// with an additional .sam extension. The unmapped alignments are collected in the file splitChrSam_unaligned.sam
// Returns the chromosome names in the order in which they occur in the sam file header
SEXP split_sam_chr(SEXP samFile, SEXP outDir)
{
if (!Rf_isString(samFile) || 1 != Rf_length(samFile)){
Rf_error("'samFile' must be character(1)");
}
if (!Rf_isString(outDir) || 1 != Rf_length(outDir)){
Rf_error("'outDir' must be character(1)");
}
const char * sam_file = Rf_translateChar(STRING_ELT(samFile, 0));
const char * out_dir = Rf_translateChar(STRING_ELT(outDir, 0));
// open the input sam file
samfile_t *fin = _bam_tryopen(sam_file, "r", NULL);
if (fin->header == 0) {
samclose(fin);
Rf_error("invalid header");
}
// remove \r from header if exists (for windows)
int j, k = 0;
for(j = 0; j<fin->header->l_text; j++){
if(fin->header->text[j] != '\r'){
fin->header->text[k++] = fin->header->text[j];
}
}
if(j != k){
fin->header->text[k] = '\0';
fin->header->l_text = (uint32_t)strlen(fin->header->text);
}
// allocate memory for a list of filehandles (n+1 because of the unaligned reads)
samfile_t **foutList = (samfile_t**)calloc((size_t)(fin->header->n_targets+1), sizeof(samfile_t*));
// open the output file handles (n+1 due to the unaligned reads)
int i;
SEXP chrNames;
PROTECT(chrNames = allocVector(STRSXP, (fin->header->n_targets+1))); // protect from garbage collector
for (i = 0; i < (fin->header->n_targets); i++) {
foutList[i] = _bam_tryopen(_assemble_file_name(out_dir,fin->header->target_name[i]), "wh", fin->header);
SET_STRING_ELT(chrNames, i, mkChar(fin->header->target_name[i]));
}
foutList[fin->header->n_targets] = _bam_tryopen(_assemble_file_name(out_dir,"splitChrSam_unaligned"), "wh", fin->header);
SET_STRING_ELT(chrNames, fin->header->n_targets, mkChar("splitChrSam_unaligned"));
// split the sam file based on chromosome
_walk_through_sam_and_split(fin,foutList);
// close all the file handles
for (i = 0; i < (fin->header->n_targets+1); i++){samclose(foutList[i]);}
samclose(fin);
UNPROTECT(1); // release
return chrNames;
}
static BAM_FILE _bamfile_open_r(SEXP filename, SEXP indexname, SEXP filemode)
{
BAM_FILE bfile = (BAM_FILE) Calloc(1, _BAM_FILE);
bfile->file = NULL;
if (0 != Rf_length(filename)) {
const char *cfile = translateChar(STRING_ELT(filename, 0));
bfile->file = _bam_tryopen(cfile, CHAR(STRING_ELT(filemode, 0)), 0);
if ((bfile->file->type & TYPE_BAM) != 1) {
samclose(bfile->file);
Free(bfile);
Rf_error("'filename' is not a BAM file\n file: %s", cfile);
}
bfile->pos0 = bam_tell(bfile->file->x.bam);
bfile->irange0 = 0;
}
bfile->index = NULL;
if (0 != Rf_length(indexname)) {
const char *cindex = translateChar(STRING_ELT(indexname, 0));
bfile->index = _bam_tryindexload(cindex);
if (NULL == bfile->index) {
samclose(bfile->file);
Free(bfile);
Rf_error("failed to open BAM index\n index: %s\n", cindex);
}
}
bfile->iter = NULL;
bfile->pbuffer = NULL;
return bfile;
}
static BAM_FILE _bamfile_open_w(SEXP file0, SEXP file1)
{
samfile_t *infile, *outfile;
BAM_FILE bfile;
if (0 == Rf_length(file1))
Rf_error("'file1' must be a character(1) path to a valid bam file");
infile = _bam_tryopen(translateChar(STRING_ELT(file1, 0)), "rb", 0);
outfile = _bam_tryopen(translateChar(STRING_ELT(file0, 0)), "wb",
infile->header);
samclose(infile);
bfile = (BAM_FILE) Calloc(1, _BAM_FILE);
bfile->file = outfile;
bfile->pos0 = bam_tell(bfile->file->x.bam);
bfile->irange0 = 0;
return bfile;
}
