int build_main (int argc, char **argv)
{
if (argc < 2)
build_usage ();
char *position;
BIGNUM capacity;
/*-------defaults for bloom filter building-------*/
int opt;
int k_mer = 0;
float error_rate = 0.0005;
char *list = NULL;
char *target_path = NULL;
char *source = NULL;
//XXX make -l and -r mutually exclusive
while ((opt = getopt (argc, argv, "e:k:o:r:l:h")) != -1)
{
switch (opt)
{
case 'e':
(optarg) && ((error_rate = atof (optarg)), 1);
break;
case 'k':
(optarg) && ((k_mer = atoi (optarg)), 1);
break;
case 'o':
(optarg) && ((target_path = optarg), 1);
break;
case 'r':
(optarg) && (source = optarg, 1);
break;
case 'l':
(optarg) && (list = optarg, 1);
break;
case 'h':
return build_usage ();
default:
printf ("Unknown option: -%c\n", (char) optopt);
return build_usage ();
}
}
if (!list && !source)
{
fprintf (stderr, "\nPlease, at least specify a reference file (-r) and an output bloom filter (-o)\n");
exit (-1);
}
if (!list)
{
#ifdef DEBUG
printf ("[bloom build]: source is %s\n", source);
printf ("[bloom build]: target is %s\n", target_path);
#endif
build (source, target_path, k_mer, error_rate, argv[0]);
}
else
{
bloom *bl_2 = NEW (bloom);
Queue *head = NEW (Queue);
Queue *tail = NEW (Queue);
head->next = tail;
F_set *File_head = NEW (F_set);
File_head = make_list (source, list);
while (File_head)
{
//map query- into memory--------------
position = mmaping (File_head->filename);
if (*position == '>')
capacity = strlen (position);
else
capacity = strlen (position) / 2;
init_bloom (bl_2, capacity, error_rate, k_mer, File_head->filename);
ref_add (bl_2, position);
save_bloom (File_head->filename, bl_2, argv[0], target_path);
bloom_destroy (bl_2);
munmap (position, strlen (position));
File_head = File_head->next;
}
}
return 0;
}
int main(int argc, char **argv)
{
uint r = ROUNDS;
uint only_crea = 0;
uint do_judy = 0;
uint seed = 0;
float scaler = 0;
uint sum = 0;
uint false_hits = 0;
uint nof_false = 0;
Pvoid_t judy = NULL;
dub_init();
PPARM_INT(seed, SEED);
PPARM_FLOAT(scaler, SCALER);
/* allow at most one false hit in 10^7 queries */
init_bloom(7, scaler);
if (getenv("ONLY_CREATE")){
dub_msg("Only encoding %u lists", r);
only_crea = 1;
}else
dub_msg("Encoding %u lists and testing %u items per list",
r, TESTMAX);
if (!only_crea && getenv("FALSE_HITS")){
dub_msg("Checking for false hits");
false_hits = 1;
}else if (getenv("JUDY")){
dub_msg("Using Judy");
do_judy = 1;
}
srand(seed);
while (r--){
uint sze = MAXSZE * (rand() / (RAND_MAX + 1.0));
u32 *p = xmalloc(sze * 4);
bloom_s *b = NULL;
uint j, k;
for (j = 0; j < sze; j++)
p[j] = 1000 * (rand() / (RAND_MAX + 1.0));
if (do_judy || false_hits){
uint tmp;
for (j = 0; j < sze; j++)
J1S(tmp, judy, p[j]);
}
if (!do_judy)
b = new_bloom(p, sze);
sum += sze;
if (only_crea)
goto next;
for (j = 1; j < TESTMAX; j++){
if (false_hits){
J1T(k, judy, j);
if (bloom_test(b, j)){
if (!k) ++nof_false;
}else
if (k){
print(p, sze);
dub_die("False negative! "
"Value %u not in bloom. "
"Bloom broken!", j);
}
}else if (do_judy){
J1T(k, judy, j);
}else
bloom_test(b, j);
}
next:
if (do_judy || false_hits){
J1FA(k, judy);
}
if (!do_judy){
free(b);
free(p);
}
}
dub_msg("In total %u items in lists", sum);
if (false_hits)
dub_msg("In total %u false hits found", nof_false);
return 0;
}