// Insert Ref with hash key Key into global Hash_Table .
// Ref represents string S .
static
void
Hash_Insert(String_Ref_t Ref, uint64 Key, char * S) {
String_Ref_t H_Ref;
char * T;
int Shift;
unsigned char Key_Check;
int64 Ct, Probe, Sub;
int i;
Sub = HASH_FUNCTION (Key);
Shift = HASH_CHECK_FUNCTION (Key);
Hash_Check_Array [Sub] |= (((Check_Vector_t) 1) << Shift);
Key_Check = KEY_CHECK_FUNCTION (Key);
Probe = PROBE_FUNCTION (Key);
Ct = 0;
do {
for (i = 0; i < Hash_Table [Sub] . Entry_Ct; i ++)
if (Hash_Table [Sub] . Check [i] == Key_Check) {
H_Ref = Hash_Table [Sub] . Entry [i];
T = Data + String_Start [getStringRefStringNum(H_Ref)] + getStringRefOffset(H_Ref);
if (strncmp (S, T, Kmer_Len) == 0) {
if (getStringRefLast(H_Ref)) {
Extra_Ref_Ct ++;
}
Next_Ref [(String_Start [getStringRefStringNum(Ref)] + getStringRefOffset(Ref)) / (HASH_KMER_SKIP + 1)] = H_Ref;
Extra_Ref_Ct ++;
setStringRefLast(Ref, TRUELY_ZERO);
Hash_Table [Sub] . Entry [i] = Ref;
if (Hash_Table [Sub] . Hits [i] < HIGHEST_KMER_LIMIT)
Hash_Table [Sub] . Hits [i] ++;
return;
}
}
if (i != Hash_Table [Sub] . Entry_Ct) {
fprintf (stderr, "i = %d Sub = " F_S64 " Entry_Ct = %d\n",
i, Sub, Hash_Table [Sub] . Entry_Ct);
}
assert (i == Hash_Table [Sub] . Entry_Ct);
if (Hash_Table [Sub] . Entry_Ct < ENTRIES_PER_BUCKET) {
setStringRefLast(Ref, TRUELY_ONE);
Hash_Table [Sub] . Entry [i] = Ref;
Hash_Table [Sub] . Check [i] = Key_Check;
Hash_Table [Sub] . Entry_Ct ++;
Hash_Entries ++;
Hash_Table [Sub] . Hits [i] = 1;
return;
}
Sub = (Sub + Probe) % HASH_TABLE_SIZE;
} while (++ Ct < HASH_TABLE_SIZE);
fprintf (stderr, "ERROR: Hash table full\n");
assert (FALSE);
}
// Add string s as an extra hash table string and return
// a single reference to the beginning of it.
static
String_Ref_t
Add_Extra_Hash_String(const char * s) {
String_Ref_t ref = 0;
String_Ref_t sub = 0;
size_t new_len;
int len;
new_len = Used_Data_Len + Kmer_Len;
if (Extra_String_Subcount < MAX_EXTRA_SUBCOUNT) {
sub = String_Ct + Extra_String_Ct - 1;
} else {
sub = String_Ct + Extra_String_Ct;
if (sub >= String_Start_Size) {
String_Start_Size *= MEMORY_EXPANSION_FACTOR;
if (sub >= String_Start_Size)
String_Start_Size = sub;
String_Start = (int64 *) safe_realloc (String_Start,
String_Start_Size * sizeof (int64));
}
String_Start [sub] = Used_Data_Len;
Extra_String_Ct ++;
Extra_String_Subcount = 0;
new_len ++;
}
if (new_len >= Extra_Data_Len) {
Extra_Data_Len = (size_t) (Extra_Data_Len * MEMORY_EXPANSION_FACTOR);
if (new_len > Extra_Data_Len)
Extra_Data_Len = new_len;
Data = (char *) safe_realloc (Data, Extra_Data_Len);
}
strncpy (Data + String_Start [sub] + Kmer_Len * Extra_String_Subcount,
s, Kmer_Len + 1);
Used_Data_Len = new_len;
setStringRefStringNum(ref, sub);
if (sub > MAX_STRING_NUM) {
fprintf (stderr, "Too many skip kmer strings for hash table.\n"
"Try skipping hopeless check (-z option)\n"
"Exiting\n");
exit (1);
}
setStringRefOffset(ref, (String_Ref_t)Extra_String_Subcount * (String_Ref_t)Kmer_Len);
assert(Extra_String_Subcount * Kmer_Len < OFFSET_MASK);
setStringRefLast(ref, TRUELY_ONE);
setStringRefEmpty(ref, TRUELY_ONE);
Extra_String_Subcount ++;
return ref;
}
// Add string s as an extra hash table string and return
// a single reference to the beginning of it.
static
String_Ref_t
Add_Extra_Hash_String(const char *s) {
String_Ref_t ref = 0;
String_Ref_t sub = 0;
int len;
uint32 new_len = Used_Data_Len + G.Kmer_Len;
if (Extra_String_Subcount < MAX_EXTRA_SUBCOUNT) {
sub = String_Ct + Extra_String_Ct - 1;
} else {
sub = String_Ct + Extra_String_Ct;
if (sub >= String_Start_Size) {
uint64 n = max(sub * 1.1, String_Start_Size * 1.5);
//fprintf(stderr, "REALLOC String_Start from "F_U64" to "F_U64"\n", String_Start_Size, n);
resizeArray(String_Start, String_Start_Size, String_Start_Size, n);
}
String_Start[sub] = Used_Data_Len;
Extra_String_Ct++;
Extra_String_Subcount = 0;
new_len++;
}
if (new_len >= Extra_Data_Len) {
uint64 n = max(new_len * 1.1, Extra_Data_Len * 1.5);
//fprintf(stderr, "REALLOC basesData from "F_U64" to "F_U64"\n", Extra_Data_Len, n);
resizeArray(basesData, Extra_Data_Len, Extra_Data_Len, n);
}
strncpy(basesData + String_Start[sub] + G.Kmer_Len * Extra_String_Subcount, s, G.Kmer_Len + 1);
Used_Data_Len = new_len;
setStringRefStringNum(ref, sub);
if (sub > MAX_STRING_NUM) {
fprintf(stderr, "Too many skip kmer strings for hash table.\n");
fprintf(stderr, "Try skipping hopeless check (-z option)\n");
fprintf(stderr, "Exiting\n");
exit (1);
}
setStringRefOffset(ref, (String_Ref_t)Extra_String_Subcount * (String_Ref_t)G.Kmer_Len);
assert(Extra_String_Subcount * G.Kmer_Len < OFFSET_MASK);
setStringRefLast(ref, (uint64)1);
setStringRefEmpty(ref, TRUELY_ONE);
Extra_String_Subcount++;
return(ref);
}
