/**
* Returns the size of the set. May be approximate.
* @arg s The set to query
* @return The size of the set.
*/
uint64_t set_size(set_t *s) {
switch (s->type) {
case EXACT:
return s->store.s.count;
case APPROX:
return ceil(hll_size(&s->store.h));
default:
abort();
}
}
END_TEST
START_TEST(test_hll_size)
{
hll_t h;
fail_unless(hll_init(10, &h) == 0);
double s = hll_size(&h);
fail_unless(s == 0);
fail_unless(hll_destroy(&h) == 0);
}
int main() {
hll_t *hll = malloc(sizeof(hll_t));
int res = hll_init(18, hll);
char line[MAXLINELEN];
double after;
while (fgets(line, MAXLINELEN, stdin)) {
hll_add(hll, line);
}
after = hll_size(hll);
printf("%.1f\n", after);
hll_destroy(hll);
return 0;
}
END_TEST
START_TEST(test_hll_add_size)
{
hll_t h;
fail_unless(hll_init(10, &h) == 0);
char buf[100];
for (int i=0; i < 100; i++) {
fail_unless(sprintf((char*)&buf, "test%d", i));
hll_add(&h, (char*)&buf);
}
double s = hll_size(&h);
fail_unless(s > 95 && s < 105);
fail_unless(hll_destroy(&h) == 0);
}
END_TEST
START_TEST(test_hll_error_bound)
{
// Precision 14 -> variance of 1%
hll_t h;
fail_unless(hll_init(14, &h) == 0);
char buf[100];
for (int i=0; i < 10000; i++) {
fail_unless(sprintf((char*)&buf, "test%d", i));
hll_add(&h, (char*)&buf);
}
// Should be within 1%
double s = hll_size(&h);
fail_unless(s > 9900 && s < 10100);
fail_unless(hll_destroy(&h) == 0);
}
CAMLprim value
caml_hll_size(value hll) {
hll_t *_hll = (hll_t*)hll;
double size = hll_size(_hll);
return caml_copy_double(size);
}
