void sort_by_name( Taxonomy& tax, bool recursive, bool case_sensitive )
{
// Make two functions for case sensitive and insensitive comparison.
auto comp_by_name_cs = []( Taxon const& lhs, Taxon const& rhs ) {
return lhs.name() < rhs.name();
};
auto comp_by_name_ci = []( Taxon const& lhs, Taxon const& rhs ) {
return utils::to_lower( lhs.name() ) < utils::to_lower( rhs.name() );
};
// Sort.
if( case_sensitive ) {
// std::sort( tax.begin(), tax.end(), comp_by_name_cs );
tax.sort( comp_by_name_cs );
} else {
// std::sort( tax.begin(), tax.end(), comp_by_name_ci );
tax.sort( comp_by_name_ci );
}
// Run recursion if necessary.
if( recursive ) {
for( auto& child : tax ) {
sort_by_name( child, true, case_sensitive );
}
}
}
// Generate Transactions and Taxonomy
//
void gen_taxrules(TaxPar &par)
{
Taxonomy *tax;
StringSet *lits;
StringSetIter *patterns;
Transaction *trans;
poisson_distribution<LINT> tlen(par.tlen - 1);
ofstream data_fp;
ofstream pat_fp;
ofstream tax_fp;
data_fp.open(data_file, ios::trunc);
pat_fp.open(pat_file, ios::trunc);
tax_fp.open(tax_file, ios::trunc);
if (data_fp.fail() || pat_fp.fail() || tax_fp.fail()) {
cerr << "Error opening output file" << endl;
exit(1);
}
// generate taxonomy and write it to file
tax = new Taxonomy(par.nitems, par.nroots, par.fanout, par.depth_ratio);
if (par.ascii)
tax->write_asc(tax_fp);
else
tax->write(tax_fp);
lits = new StringSet(par.nitems, par.lits, tax);
par.write(pat_fp);
lits->display(pat_fp);
patterns = new StringSetIter(*lits);
for (LINT i = 0; i < par.ntrans; i ++)
{
trans = mk_tran(*patterns, tlen(generator) + 1, tax);
if (par.ascii)
trans->write_asc(data_fp);
else
trans->write(data_fp);
delete trans;
delete trans;
}
data_fp.close();
pat_fp.close();
tax_fp.close();
}
bool has_unique_ids( Taxonomy const& tax )
{
std::unordered_set<std::string> ids;
bool has_duplicates = false;
auto collect_and_check = [&]( Taxon const& tax ){
if( ids.count( tax.id() ) > 0 ) {
has_duplicates = true;
return;
}
ids.insert( tax.id() );
};
preorder_for_each( tax, collect_and_check );
return ! has_duplicates;
}
size_t total_taxa_count( Taxonomy const& tax )
{
size_t count = tax.size();
for( auto const& t : tax ) {
count += total_taxa_count( t );
}
return count;
}
void remove_taxa_at_level( Taxonomy& tax, size_t level )
{
// Recursive implementation, because we are lazy.
if( level == 0 ) {
tax.clear_children();
} else {
for( auto& c : tax ) {
remove_taxa_at_level( c, level - 1 );
}
}
}
std::vector< size_t > taxa_count_levels( Taxonomy const& tax )
{
if( tax.size() == 0 ) {
return std::vector< size_t >();
}
std::vector< size_t > result( 1, 0 );
result[ 0 ] = tax.size();
for( auto const& child : tax ) {
auto cres = taxa_count_levels( child );
if( result.size() < cres.size() + 1 ) {
result.resize( cres.size() + 1, 0 );
}
for( size_t i = 0; i < cres.size(); ++i ) {
result[ i+1 ] += cres[ i ];
}
}
return result;
}
size_t taxa_count_at_level( Taxonomy const& tax, size_t level )
{
// Recursive implementation, because we are lazy.
size_t count = 0;
if( level == 0 ) {
count += tax.size();
} else {
for( auto& c : tax ) {
count += taxa_count_at_level( c, level - 1 );
}
}
return count;
}
// Generate Transactions and Taxonomy
//
void gen_taxrules(TaxPar &par)
{
Taxonomy *tax;
StringSet *lits;
StringSetIter *patterns;
Transaction *trans;
PoissonDist *tlen;
ofstream data_fp;
ofstream pat_fp;
ofstream tax_fp;
ofstream conf_fp; //added by MJZaki
data_fp.open(data_file);
pat_fp.open(pat_file);
tax_fp.open(tax_file);
conf_fp.open(conf_file); //added by MJZaki
if (data_fp.fail() || pat_fp.fail() || tax_fp.fail() || conf_fp.fail())
{
cerr << "Error opening output file" << endl;
exit(1);
}
// generate taxonomy and write it to file
tax = new Taxonomy(par.nitems, par.nroots, par.fanout, par.depth_ratio);
if (par.ascii)
tax->write_asc(tax_fp);
else
tax->write(tax_fp);
tlen = new PoissonDist(par.tlen - 1);
lits = new StringSet(par.nitems, par.lits, tax);
par.write(pat_fp);
lits->display(pat_fp);
patterns = new StringSetIter(*lits);
LINT NTRANS = 0;
for (LINT i = 0; i < par.ntrans; i++)
{
trans = mk_tran(*patterns, (*tlen)() + 1, par.lits.npats, tax);
if (par.ascii)
trans->write_asc(data_fp);
else
trans->write(data_fp);
if (trans->size() > 0)
NTRANS++;//added by MJZaki: repeat if trans empty
else
i--;
delete trans;
}
data_fp.close();
pat_fp.close();
tax_fp.close();
//added by MJZaki
if (par.ascii)
{
conf_fp << NTRANS << "\n";
conf_fp << par.nitems << "\n";
conf_fp << par.tlen << "\n";
}
else
{
conf_fp.write((char *) &NTRANS, sizeof(LINT));
conf_fp.write((char *) &par.nitems, sizeof(LINT));
int t = (int) par.tlen;
conf_fp.write((char *) &t, sizeof(LINT));
}
conf_fp.close();
}
