// generate a ground substitution for a clause
int
groundSubstitutions(const Clause &clause, Substitutions &subs)
{
// clear substitutions
subs.clear();
// get a string representation
String clstring = (String)clause;
// list of variables
Map<String, String> variables;
// find variables in clause
StringTokens clst(clstring, " \t");
for ( ; !clst.done(); clst++)
{
if ((clst()(0,2) == String("_V")) ||
(clst()(0,3) == String("_RV")))
{
String variable(clst());
if (!variables.isInMap(variable))
{
variables[variable] =
uniqueName("_CONST");
}
}
}
// generate substitutions now
MapIterator<String, String> varsIter(variables);
for ( ; !varsIter.done(); varsIter++)
{
Substitution sub(Terms(varsIter.data()), varsIter.key());
subs = subs*sub;
}
// all done
return(OK);
}
int main(void)
{
clst("lower123", "lowerabc", '\0');
clst("higher234", "higher123", '\0');
clst("equal", "equal", '\0');
clst("equal", "equallong", '\0');
clst("equallong", "equal", '\0');
clst("lower1", "lower2", 'w'); // will compare equal
return 0;
}
/** Function make_cluster()
*
* Given a list of fragments denoted by seeds, make pairwise comparison
* and clustering conforming max_mismatch criteria
*
* Output: clusters in 2d vector format, where each row of the vector
* stores the clustered fragment IDs.
*/
void make_cluster (iivec_t& clusters, const ii64vec_t& list_seeds,
const ivec_t& init_cluster, int max_mismatch, const ivec_t& uf_clst) {
if (list_seeds.size() == 0) {
abording ("DuplRm.cpp -- make_cluster(): SC failed");
}
//--------- union find: (1) initialize the cluster ---------
int sz = init_cluster.size();
bvec_t visited (sz, false);
ivec_t clst (sz);
for (int i = 0; i < sz; ++ i) clst[i] = i;
//--------- pairwise comparison ---------
for (int i = 0; i < sz - 1; ++ i) {
if (visited[i]) continue; // to speed up
int idx_i = init_cluster[i];
for (int j = i + 1; j < sz; ++ j) {
if (visited[j]) continue; // to speed up, avoid of comparison
// if this is already clustered
int idx_j = init_cluster[j];
// check global uf structure according to fragID
int root_uf_i = uf_clsfind ((int) list_seeds[idx_i].back(), uf_clst),
root_uf_j = uf_clsfind ((int) list_seeds[idx_j].back(), uf_clst);
if (root_uf_i != root_uf_j) {
int root_i = uf_find (i, clst),
root_j = uf_find (j, clst);
if (root_i != root_j) {
if (is_similar (list_seeds[idx_i], list_seeds[idx_j],
max_mismatch)) {
clst[root_j] = root_i;
visited[j] = true;
}
}
} // if
} // for (int j = i + 1
} // for (int i = 0
//----- generate final cluster { clusterID --> fragment IDs } ------
std::map<int, ivec_t> clstID_fragIDs;
std::map<int, ivec_t>::iterator it;
for (int i = 0; i < sz; ++ i) {
int idx_i = init_cluster[i];
int fragID = list_seeds[idx_i].back();
int root_i = uf_clsfind (i, clst);
it = clstID_fragIDs.find (root_i);
if (it != clstID_fragIDs.end()) it->second.push_back(fragID);
else clstID_fragIDs[root_i] = ivec_t (1, fragID);
} // for (int i = 0
// go through the map and produce clusters in sorted vector format
for (it = clstID_fragIDs.begin(); it != clstID_fragIDs.end(); ++ it) {
if (it->second.size() > 1) {
std::sort (it->second.begin(), it->second.end());
clusters.push_back(it->second);
}
} // for (it
} // make_cluster
