static unsigned dbget_raw(struct fheader * h, unsigned busy){
uint32_t idx;
if( h->wbit){
idx = search(busy,blist[h->rank],cnk[h->rank]);
return twobit_get(h->data,idx);
}
if (h->sorted){
idx = search(busy,(uint32_t *)(h->data),h->count0);
if (idx != -1)
return r0[h->value];
idx = search(busy,(uint32_t *)(h->data)+h->count0,h->count1);
if (idx != -1)
return r1[h->value];
}else {
for(idx=0;idx<h->count0;idx++)
if(((uint32_t*)(h->data))[idx] == busy)
return r0[h->value];
for(;idx<h->count0+h->count1;idx++)
if(((uint32_t*)(h->data))[idx] == busy)
return r1[h->value];
}
if (h->full){
return h->value;
}
error("Not in db");
}
void seqenc_encode_alignment(
seqenc_t* E,
uint32_t spos, uint8_t strand,
const unsigned char* query_str,
const twobit_t* query)
{
size_t qlen = twobit_len(query);
size_t slen = E->supercontig_motif.n;
assert(spos < slen);
dist2_encode(E->ac, &E->d_type, SEQENC_TYPE_ALIGNMENT);
/* encode N mask */
size_t i;
reserve_nmask(E, qlen);
for (i = 0; i < qlen; ++i) {
dist2_encode(E->ac, &E->d_nmask[i],
query_str[i] == 'N' ? 1 : 0);
}
dist2_encode(E->ac, &E->d_aln_strand, strand);
uint32_enc_encode(E->ac, &E->d_contig_off, spos);
kmer_t u;
if (strand) {
for (i = 0; i < qlen; ++i) {
if (query_str[i] == 'N') continue;
u = kmer_comp1(twobit_get(query, i));
cond_dist4_encode(E->ac, &E->supercontig_motif, slen - (spos + i) - 1, u);
}
}
else {
for (i = 0; i < qlen; ++i) {
if (query_str[i] == 'N') continue;
u = twobit_get(query, i);
cond_dist4_encode(E->ac, &E->supercontig_motif, spos + i, u);
}
}
}
void seqenc_set_supercontig(seqenc_t* E, const twobit_t* supercontig)
{
size_t len = twobit_len(supercontig);
if (len == 0) return;
cond_dist4_init(&E->supercontig_motif, len);
cond_dist4_set_update_rate(&E->supercontig_motif, motif_update_rate);
size_t i;
kmer_t u;
for (i = 0; i < len; ++i) {
u = twobit_get(supercontig, i);
E->supercontig_motif.xss[i].xs[u].count = contig_motif_prior;
dist4_update(&E->supercontig_motif.xss[i]);
}
}
void seqenc_encode_twobit_seq(seqenc_t* E, const unsigned char* x_str, const twobit_t* x)
{
dist2_encode(E->ac, &E->d_type, SEQENC_TYPE_SEQUENCE);
size_t n = twobit_len(x);
if (n == 0) return;
kmer_t uv;
uint32_t ctx = 0;
size_t i;
for (i = 0; i < n - 1 && i / 2 < prefix_len; i += 2) {
uv = (twobit_get(x, i) << 2) | twobit_get(x, i + 1);
cond_dist16_encode(E->ac, &E->cs0[i/2], ctx, uv);
ctx = ((ctx << 4) | uv) & E->ctx_mask;
}
for (; i < n - 1; i += 2) {
uv = (twobit_get(x, i) << 2) | twobit_get(x, i + 1);
cond_dist16_encode(E->ac, &E->cs, ctx, uv);
ctx = ((ctx << 4) | uv) & E->ctx_mask;
}
/* handle odd read lengths */
if (i < n) {
uv = twobit_get(x, i);
cond_dist16_encode(E->ac, &E->cs, ctx, uv);
}
/* encode N mask */
reserve_nmask(E, n);
for (i = 0; i < n; ++i) {
dist2_encode(E->ac, &E->d_nmask[i],
x_str[i] == 'N' ? 1 : 0);
}
}
int main(int argc, char ** argv){
uint32_t x,b,w,d;
unsigned char * array;
int rank;
int idx = 0;
if(argc!=4) panic();
b = getarg(1);
w = getarg(2);
d = getarg(3);
rank = _popc(b);
array = malloc_file(ARRAY_SIZE_S(rank),FMODE_RO,DATA_FORMAT,rank);
// search index
for(x=ALLONE(rank);x!=b;x=_permut(x))
idx++;
if ( blist_get(b) != idx )
error("What's f***a with blist\n");
//
int res = twobit_get(array+(uint64_t)((w<<rank)|d)*cnk(32,rank)/4,idx);
printf("%08X %X %X = %d\n",b,w,d,res);
return 0;
}
static void seqenc_decode_reference_alignment(seqenc_t* E, short_read_t* r, size_t seqlen)
{
const twobit_t* refseq = seqmap_get(E->ref, (const char*) r->seqname.s);
if (refseq == NULL) {
quip_error(
"A read was aligned to sequence %s, which was not found in the reference.",
r->seqname.s);
}
str_reserve(&r->seq, seqlen + 1);
r->seq.n = 0;
/* decode N mask */
size_t i;
reserve_nmask(E, seqlen);
memset(r->seq.s, '\0', seqlen + 1);
for (i = 0; i < seqlen; ++i) {
if (dist2_decode(E->ac, &E->d_nmask[i])) r->seq.s[i] = 'N';
}
uint32_t ref_pos = r->pos;
uint32_t read_pos = 0;
i = 0; /* cigar operation */
size_t j; /* position within the cigar op */
kmer_t y; /* reference nucleotide */
for (i = 0; i < r->cigar.n; ++i) {
switch (r->cigar.ops[i]) {
case BAM_CEQUAL:
case BAM_CDIFF:
case BAM_CMATCH:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos, ++ref_pos) {
if (r->seq.s[read_pos] == 'N') continue;
if (dist2_decode(E->ac, &E->d_ref_match) == SEQENC_REF_MATCH) {
y = twobit_get(refseq, ref_pos);
r->seq.s[read_pos] = kmertochar[y];
}
else {
r->seq.s[read_pos] = kmertochar[dist4_decode(E->ac, &E->d_ref_ins_nuc)];
}
}
break;
case BAM_CINS:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos) {
if (r->seq.s[read_pos] == 'N') continue;
r->seq.s[read_pos] = kmertochar[dist4_decode(E->ac, &E->d_ref_ins_nuc)];
}
break;
case BAM_CDEL:
ref_pos += r->cigar.lens[i];
break;
case BAM_CREF_SKIP:
ref_pos += r->cigar.lens[i];
break;
case BAM_CSOFT_CLIP:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos) {
if (r->seq.s[read_pos] == 'N') continue;
r->seq.s[read_pos] = kmertochar[dist4_decode(E->ac, &E->d_ref_ins_nuc)];
}
break;
case BAM_CHARD_CLIP:
ref_pos += r->cigar.lens[i];
break;
case BAM_CPAD:
quip_error("Unsupported cigar operation.");
break;
}
}
r->seq.s[seqlen] = '\0';
r->seq.n = seqlen;
if (read_pos != seqlen) {
quip_error("Cigar operations do not account for full read length.");
}
if (r->strand) str_revcomp(r->seq.s, r->seq.n);
}
void seqenc_encode_reference_alignment(
seqenc_t* E, const short_read_t* r)
{
const twobit_t* refseq = seqmap_get(E->ref, (const char*) r->seqname.s);
if (refseq == NULL) {
quip_error(
"A read was aligned to sequence %s, which was not found in the reference.\n",
r->seqname.s);
}
str_copy(&E->tmpseq, &r->seq);
if (r->strand) {
str_revcomp(E->tmpseq.s, E->tmpseq.n);
}
/* encode N mask */
size_t i;
reserve_nmask(E, r->seq.n);
for (i = 0; i < r->seq.n; ++i) {
dist2_encode(E->ac, &E->d_nmask[i],
E->tmpseq.s[i] == 'N' ? 1 : 0);
}
uint32_t ref_pos = r->pos;
uint32_t read_pos = 0;
i = 0; /* cigar operation */
size_t j; /* position within the cigar op */
kmer_t x; /* read nucleotide */
kmer_t y; /* reference nucleotide */
for (i = 0; i < r->cigar.n; ++i) {
switch (r->cigar.ops[i]) {
case BAM_CEQUAL:
case BAM_CDIFF:
case BAM_CMATCH:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos, ++ref_pos) {
if (E->tmpseq.s[read_pos] == 'N') continue;
x = chartokmer[E->tmpseq.s[read_pos]];
y = twobit_get(refseq, ref_pos);
if (x == y) {
dist2_encode(E->ac, &E->d_ref_match, SEQENC_REF_MATCH);
}
else {
dist2_encode(E->ac, &E->d_ref_match, SEQENC_REF_MISMATCH);
dist4_encode(E->ac, &E->d_ref_ins_nuc, chartokmer[E->tmpseq.s[read_pos]]);
}
}
break;
case BAM_CINS:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos) {
if (E->tmpseq.s[read_pos] == 'N') continue;
dist4_encode(E->ac, &E->d_ref_ins_nuc, chartokmer[E->tmpseq.s[read_pos]]);
}
break;
case BAM_CDEL:
ref_pos += r->cigar.lens[i];
break;
case BAM_CREF_SKIP:
ref_pos += r->cigar.lens[i];
break;
case BAM_CSOFT_CLIP:
for (j = 0; j < r->cigar.lens[i]; ++j, ++read_pos) {
if (E->tmpseq.s[read_pos] == 'N') continue;
dist4_encode(E->ac, &E->d_ref_ins_nuc, chartokmer[E->tmpseq.s[read_pos]]);
}
break;
case BAM_CHARD_CLIP:
ref_pos += r->cigar.lens[i];
break;
case BAM_CPAD:
quip_error("Cigar PAD operation is unsupported.");
break;
}
}
if (read_pos != r->seq.n) {
quip_error("Cigar operations do not account for full read length.");
}
}
