void command_line(TaxPar &par)
{
par.calc_values();
cerr << "Command Line Options:" << endl;
cerr << " -ntrans number_of_transactions_in_000s (default: "
<< par.ntrans/1000 << ")" << endl;
cerr << " -tlen avg_items_per_transaction (default: " << par.tlen << ")" << endl;
cerr << " -nitems number_of_different_items_in_000s (default: "
<< par.nitems/1000 << ")" << endl;
cerr << " -nroots number_of_roots (default: " << par.nroots << ")" << endl;
cerr << " -nlevels number_of_different_levels (default: " << par.nlevels
<< ")" << endl;
cerr << " -fanout average_fanout (default: " << par.fanout << ")" << endl;
cerr << " -depth affects_average_depth_of_items_in_itemsets (default: "
<< par.depth_ratio << ")" << endl;
cerr << endl;
cerr << " -npats number_of_patterns (default: " << par.lits.npats << ")" << endl;
cerr << " -patlen avg_length_of_maximal_pattern (default: "
<< par.lits.patlen << ")" << endl;
cerr << " -corr correlation_between_patterns (default: " << par.lits.corr
<< ")" << endl;
cerr << " -conf avg_confidence_in_a_rule (default: " << par.lits.conf
<< ")" << endl;
cerr << endl;
cerr << " -fname <filename> (write to filename.data and filename.pat)" << endl;
cerr << " -ascii (Write data in ASCII format; default: " << (par.ascii? "True": "False") << ")" << endl;
cerr << " -randseed # (reset seed used generate to x-acts; must be negative)" << endl;
cerr << " -version (to print out version info)" << endl;
exit(EXIT_FAILURE);
}
// 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();
}
void get_args(TaxPar &par, int argc, char **argv)
{
LINT arg_pos = 2;
strcpy(data_file, "data");
strcpy(pat_file, "pat");
strcpy(tax_file, "tax");
while (arg_pos < argc)
{
if (strcmp(argv[arg_pos], "-ntrans") == 0) {
// g++ LINT !!
par.ntrans = LINT (1000 * atof(argv[++arg_pos]));
cat_fname(".ntrans_", argv[arg_pos]);
arg_pos++;
if (par.ntrans < 1) exit_fail_with_msg("ntrans must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-tlen") == 0) {
par.tlen = atof(argv[++arg_pos]);
cat_fname(".tlen_", argv[arg_pos]);
arg_pos++;
if (par.tlen < 1) exit_fail_with_msg("tlen must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-nitems") == 0) {
// g++ the LINT worm is back again!
par.nitems = LINT (1000 * atof(argv[++arg_pos]));
cat_fname(".nitems_", argv[arg_pos]);
arg_pos++;
if (par.nitems < 1) exit_fail_with_msg("nitems must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-nroots") == 0) {
par.nroots = atoi(argv[++arg_pos]);
cat_fname(".nroots_", argv[arg_pos]);
arg_pos++;
if (par.nroots < 1) exit_fail_with_msg("nroots must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-nlevels") == 0) {
par.nlevels = atof(argv[++arg_pos]);
cat_fname(".nlevels_", argv[arg_pos]);
arg_pos++;
if (par.nlevels < 1) exit_fail_with_msg("nlevels must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-fanout") == 0) {
par.fanout = atof(argv[++arg_pos]);
cat_fname(".fanout_", argv[arg_pos]);
arg_pos++;
if (par.fanout < 1) exit_fail_with_msg("fanout must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-depth") == 0) {
par.depth_ratio = atof(argv[++arg_pos]);
cat_fname(".depth_", argv[arg_pos]);
arg_pos++;
if (par.depth_ratio <= 0) exit_fail_with_msg("fanout must be > 0");
continue;
}
else if (strcmp(argv[arg_pos], "-npats") == 0) {
par.lits.npats = atoi(argv[++arg_pos]);
cat_fname(".npats_", argv[arg_pos]);
arg_pos++;
if (par.lits.npats < 1) exit_fail_with_msg("npats must be >= 1");
continue;
}
else if (strcmp(argv[arg_pos], "-patlen") == 0) {
par.lits.patlen = atof(argv[++arg_pos]);
cat_fname(".patlen_", argv[arg_pos]);
arg_pos++;
if (par.lits.patlen <= 0) exit_fail_with_msg("patlen must be > 0");
continue;
}
else if (strcmp(argv[arg_pos], "-corr") == 0) {
par.lits.corr = atof(argv[++arg_pos]);
cat_fname(".corr_", argv[arg_pos]);
arg_pos++;
continue;
}
else if (strcmp(argv[arg_pos], "-conf") == 0) {
par.lits.conf = atof(argv[++arg_pos]);
cat_fname(".conf_", argv[arg_pos]);
arg_pos++;
if (par.lits.conf > 1 || par.lits.conf < 0)
exit_fail_with_msg("conf must be between 0 and 1");
continue;
}
else if (strcmp(argv[arg_pos], "-fname") == 0) {
strcpy(data_file, argv[++arg_pos]);
strcat(data_file, ".data");
strcpy(pat_file, argv[arg_pos]);
strcat(pat_file, ".pat");
strcpy(tax_file, argv[arg_pos++]);
strcat(tax_file, ".tax");
userfile = true;
continue;
}
else if (strcmp(argv[arg_pos], "-ascii") == 0) {
par.ascii = true;
cat_fname(".ascii", "");
arg_pos++;
continue;
}
else if (strcmp(argv[arg_pos], "-randseed") == 0) {
par.seed = atoi(argv[++arg_pos]);
arg_pos++;
if (par.seed >= 0)
exit_fail_with_msg("randseed must be negative");
continue;
}
else if (strcmp(argv[arg_pos], "-version") == 0) {
cout << VERSION << endl;
exit(EXIT_SUCCESS);
}
else {
command_line(par);
}
} // end while
par.calc_values();
}
// 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();
}