int main(int argc, char* argv[])
{
std::vector<std::string> word_list;
std::deque<std::string> outliers;
if (!load_word_list(argc,argv,word_list))
{
return 1;
}
generate_outliers(word_list,outliers);
unsigned int random_seed = 0xA57EC3B2;
const double desired_probability_of_false_positive = 1.0 / word_list.size();
compressible_bloom_filter filter(word_list.size(),desired_probability_of_false_positive,random_seed);
filter.insert(word_list.begin(),word_list.end());
std::cout << "Filter Size\tEFPP \tOFPP \tDiff" << std::endl;
while (filter.size() > 1)
{
std::vector<std::string>::iterator it = filter.contains_all(word_list.begin(),word_list.end());
if (word_list.end() != it)
{
std::cout << "ERROR: key not found in bloom filter! =>" << (*it) << std::endl;
return 1;
}
std::size_t total_false_positive = 0;
for (std::deque<std::string>::iterator it = outliers.begin(); it != outliers.end(); ++it)
{
if (filter.contains(*it)) ++total_false_positive;
}
double pfp = total_false_positive / (1.0 * outliers.size());
printf("%11llu\t%8.7f\t%8.7f\t%8.6f\n",
static_cast<unsigned long long>(filter.size()),
filter.effective_fpp(),
pfp,
100.0 * (pfp / filter.effective_fpp()));
if (!filter.compress(5.0))
{
std::cout << "Filter cannot be compressed any further." << std::endl;
break;
}
}
return 0;
}
int main(int argc, char* argv[])
{
std::vector<std::string> word_list;
std::deque<std::string> outliers;
if (!load_word_list(argc,argv,word_list))
{
return 1;
}
generate_outliers(word_list,outliers);
unsigned int random_seed = 0;
std::size_t word_list_storage_size = 0;
for(unsigned int i = 0; i < word_list.size(); ++i)
{
word_list_storage_size += word_list[i].size();
}
std::size_t total_number_of_queries = 0;
const double desired_probability_of_false_positive = 1.0 / word_list.size();
printf("Round\tQueries \tFPQ \tIPFP \tPFP \tDPFP \tTvD \n");
while(random_seed < 1000)
{
bloom_filter filter(word_list.size(),desired_probability_of_false_positive,random_seed++);
filter.insert(word_list.begin(),word_list.end());
std::vector<std::string>::iterator it = filter.contains_all(word_list.begin(),word_list.end());
if (word_list.end() != it)
{
std::cout << "ERROR: key not found! =>" << (*it) << std::endl;
return 1;
}
std::size_t total_false_positive = 0;
for(std::deque<std::string>::iterator it = outliers.begin(); it != outliers.end(); ++it)
{
if (filter.contains(*it))
{
++total_false_positive;
}
}
total_number_of_queries += (outliers.size() + word_list.size());
// Overall false positive probability
double pfp = total_false_positive / (1.0 * outliers.size());
printf("%10llu\t%10llu\t%6llu\t%8.7f\t%8.7f\t%8.6f\t%8.6f\n",
static_cast<unsigned long long>(random_seed),
static_cast<unsigned long long>(total_number_of_queries),
static_cast<unsigned long long>(total_false_positive),
desired_probability_of_false_positive,
pfp,
(100.0 * pfp) / desired_probability_of_false_positive,
(100.0 * filter.size()) / (bits_per_char * word_list_storage_size));
}
return 0;
}
