/*************************************************
Function:
lastKmer
Description:
Searches the node that a node's kmer-edge end with.
Input:
1. ht: the graph hashtable
2. K_size: kmer size
3. node: the node whose kmer-edge will be searched
4. edge: the kmer-edge
5. is_left: whether the kmer-edge on the node's left side
Output:
1. smaller: whether the searched result, a kmer is smaller than its reversed complement
Return:
A pointer to the found node.
Null if not found.
*************************************************/
static bucket2 * lastKmer ( hashtable2 * ht, int K_size, bucket2 * node, edge_node * edge, int is_left, int & smaller ) //NEW
{
if ( !node || !edge ) { return NULL; }
kmer_t2 t_kmer, f_kmer;
t_kmer = node->kmer_t2;
kmer_t2 edge_seq;
memset ( edge_seq.kmer, 0, sizeof ( edge_seq ) );
( edge_seq.kmer ) [sizeof ( edge_seq ) / sizeof ( uint64_t ) - 1] = edge->edge;
int edge_len = edge->len + 1;
if ( edge_len > K_size )
{
fprintf ( stderr, "ERROR: g value should be no great than kmer size!\n" );
exit ( -1 );
}
kmer_t2 KMER_FILTER;
initKmerFilter ( K_size, &KMER_FILTER );
if ( is_left ) //left edge
{
kmerMoveRight ( &t_kmer, edge_len );
kmerMoveLeft ( &edge_seq, K_size - edge_len );
kmerOr ( &t_kmer, &edge_seq );
kmerAnd ( &t_kmer, &KMER_FILTER );
}
else
{
kmerMoveLeft ( &t_kmer, edge_len );
kmerOr ( &t_kmer, &edge_seq );
kmerAnd ( &t_kmer, &KMER_FILTER );
}
f_kmer = t_kmer;
reverseCompKmer ( &f_kmer, K_size );
if ( kmerCompare ( &t_kmer, &f_kmer ) > 0 )
{
t_kmer = f_kmer;
smaller = 0;
}
else { smaller = 1; }
return search_kmer ( ht, &t_kmer );
}
static string stack2string ( hashtable2 * ht, int K_size, list<stacked_node2 *> & stack )
{
static size_t call_times;
call_times++;
string full_edge;
stacked_node2 * t_stack_node = stack.front();
char tmp[1024];
uint64_t bits[2];
kmer_t2 tmp_kmer = ( t_stack_node->node->kmer_t2 );
if ( t_stack_node->is_left )
{
reverseCompKmer ( &tmp_kmer, K_size );
}
else
{
}
bitsarr2str ( tmp_kmer.kmer, K_size, tmp, sizeof ( kmer_t2 ) / sizeof ( uint64_t ) );
full_edge.append ( tmp ); //put first node
while ( t_stack_node )
{
if ( t_stack_node->edge )
{
if ( t_stack_node->is_left )
{
bits[0] = get_rev_comp_seq ( t_stack_node->edge->edge, t_stack_node->edge->len + 1 );
bitsarr2str ( bits, t_stack_node->edge->len + 1, tmp, 1 );
full_edge.append ( tmp );
}
else
{
bits[0] = t_stack_node->edge->edge;
bitsarr2str ( bits, t_stack_node->edge->len + 1, tmp, 1 );
full_edge.append ( tmp );
}
}
t_stack_node = t_stack_node->next;
}
return full_edge;
}
/*************************************************
Function:
process_edge
Description:
It builds vetexes from one or part of one edge sequence.
Input:
1. v_ht: hashtable
2. K_size: kmer size
3. seq: edge sequence
4. len: edge length
5. type: 1: process head and tail; 2: process head ; 3:process tail
6. edge_id: edge id
7. bal_edge: 0:palindrome 1:else
Output:
None.
Return:
None.
*************************************************/
static void process_edge ( vertex_hash2 * v_ht, int K_size, char * seq, int len, int type, size_t edge_id, bool bal_edge )
{
kmer_t2 vertex_kmer;
kmer_t2 edge_kmer;
vertex2 * v_tmp;
edge_starter2 * e_tmp;
int is_found;
bool is_left;
int edge_kmer_len;
switch ( type )
{
case 1: //process all ..
//process the head
get_kmer_from_seq ( seq, len, K_size, 0, &vertex_kmer );
if ( len <= K_size + gap ) //get the last kmer
{
get_kmer_from_seq ( seq, len, K_size, len - K_size, &edge_kmer );
edge_kmer_len = len - K_size;
}
else
{
//get_kmer_from_seq(seq, len, K_size, K_size,&edge_kmer);
get_kmer_from_seq ( seq, len, K_size, gap, &edge_kmer );
edge_kmer_len = gap;
}
is_left = 0;//right
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id );
reverseCompKmer ( &vertex_kmer, K_size );
reverseCompKmer ( &edge_kmer, K_size );
is_left = 1;//left
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id + bal_edge );
//process the tail
get_kmer_from_seq ( seq, len, K_size, len - K_size, &vertex_kmer );
if ( len <= K_size + gap ) //get the first kmer
{
get_kmer_from_seq ( seq, len, K_size, 0, &edge_kmer );
edge_kmer_len = len - K_size;
}
else
{
get_kmer_from_seq ( seq, len, K_size, len - K_size - gap, &edge_kmer );
edge_kmer_len = gap;
}
is_left = 1;
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id );
reverseCompKmer ( &vertex_kmer, K_size );
reverseCompKmer ( &edge_kmer, K_size );
is_left = 0;//right
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id + bal_edge );
break;
case 2:
//process only the head
get_kmer_from_seq ( seq, len, K_size, 0, &vertex_kmer );
if ( len <= K_size + gap )
{
get_kmer_from_seq ( seq, len, K_size, len - K_size, &edge_kmer );
edge_kmer_len = len - K_size;
}
else
{
get_kmer_from_seq ( seq, len, K_size, gap, &edge_kmer );
edge_kmer_len = gap;
}
is_left = 0;//right
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id );
reverseCompKmer ( &vertex_kmer, K_size );
reverseCompKmer ( &edge_kmer, K_size );
is_left = 1;//left
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id + bal_edge );
break;
case 3:
//process only the tail
get_kmer_from_seq ( seq, len, K_size, len - K_size, &vertex_kmer );
if ( len <= K_size + gap )
{
get_kmer_from_seq ( seq, len, K_size, 0, &edge_kmer );
edge_kmer_len = len - K_size;
}
else
{
get_kmer_from_seq ( seq, len, K_size, len - K_size - gap, &edge_kmer );
edge_kmer_len = gap;
}
is_left = 1;
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id );
reverseCompKmer ( &vertex_kmer, K_size );
reverseCompKmer ( &edge_kmer, K_size );
is_left = 0;//right
v_tmp = put_vertex ( v_ht, vertex_kmer, is_found );
put_edge ( v_tmp, edge_kmer, is_left, edge_kmer_len, edge_id + bal_edge );
break;
default:
fprintf ( stderr, "ERROR: wrong process type in process_edge()\n" );
exit ( 1 );
}
}