void MangoSubstructure::_correctQueryStereo (QueryMolecule &query)
{
// Remove stereobond marks that are connected with R-groups
for (int v = query.vertexBegin();
v != query.vertexEnd();
v = query.vertexNext(v))
{
if (!query.isRSite(v))
continue;
const Vertex &vertex = query.getVertex(v);
for (int nei = vertex.neiBegin();
nei != vertex.neiEnd();
nei = vertex.neiNext(nei))
{
int edge = vertex.neiEdge(nei);
if (query.cis_trans.getParity(edge) != 0)
query.cis_trans.setParity(edge, 0);
}
}
MoleculeRGroups &rgroups = query.rgroups;
int n_rgroups = rgroups.getRGroupCount();
for (int i = 1; i <= n_rgroups; i++)
{
PtrPool<BaseMolecule> &frags = rgroups.getRGroup(i).fragments;
for (int j = frags.begin(); j != frags.end(); j = frags.next(j))
{
QueryMolecule &fragment = frags[j]->asQueryMolecule();
_correctQueryStereo(fragment);
}
}
}
void MoleculePiSystemsMatcher::_markMappedPiSystems (QueryMolecule &query,
const int *mapping)
{
for (int qv = query.vertexBegin();
qv != query.vertexEnd();
qv = query.vertexNext(qv))
{
int v = mapping[qv];
if (v < 0)
continue; // Such vertex must be ignored
int pi_system_idx = _atom_pi_system_idx[v];
if (pi_system_idx == _NOT_IN_PI_SYSTEM)
continue;
if (!_pi_systems[pi_system_idx].initialized)
_extractPiSystem(pi_system_idx);
_Pi_System &pi_system = _pi_systems[pi_system_idx];
if (!pi_system.pi_system_mapped)
{
pi_system.pi_system_mapped = true;
pi_system.localizer->unfixAll();
}
}
}
bool MoleculePiSystemsMatcher::_fixAtoms (QueryMolecule &query, const int *mapping)
{
// Fix charges
for (int qv = query.vertexBegin();
qv != query.vertexEnd();
qv = query.vertexNext(qv))
{
int v = mapping[qv];
if (v < 0)
continue; // Such vertex must be ignored
int pi_system_idx = _atom_pi_system_idx[v];
if (pi_system_idx == _NOT_IN_PI_SYSTEM)
continue;
_Pi_System &pi_system = _pi_systems[pi_system_idx];
QueryMolecule::Atom &qatom = query.getAtom(qv);
int pv = pi_system.inv_mapping[v];
int charge = query.getAtomCharge(qv);
if (charge != CHARGE_UNKNOWN)
{
bool ret = pi_system.localizer->fixAtomCharge(pv, charge);
if (!ret)
return false;
}
else if (qatom.hasConstraint(QueryMolecule::ATOM_CHARGE))
throw Error("Unsupported atom charge specified");
int valence = query.getExplicitValence(qv);
if (valence != -1)
{
bool ret = pi_system.localizer->fixAtomConnectivity(pv, valence);
if (!ret)
return false;
}
else if (qatom.hasConstraint(QueryMolecule::ATOM_VALENCE))
throw Error("Unsupported atom charge specified");
}
return true;
}
bool QueryMoleculeAromatizer::_aromatizeBonds (QueryMolecule &mol, int additional_atom, const AromaticityOptions &options)
{
bool aromatized = false;
// Mark edges that can be aromatic in some matching
aromatized |= _aromatizeBondsFuzzy(mol, options);
// Aromatize all aromatic cycles
aromatized |= _aromatizeBondsExact(mol, options);
MoleculeRGroups &rgroups = mol.rgroups;
int n_rgroups = rgroups.getRGroupCount();
// Check if r-groups are attached with single bonds
QS_DEF(Array<bool>, rgroups_attached_single);
rgroups_attached_single.clear();
for (int v = mol.vertexBegin(); v != mol.vertexEnd(); v = mol.vertexNext(v))
{
if (v == additional_atom)
continue;
if (mol.isRSite(v))
{
// Check if neighbor bonds are single
const Vertex &vertex = mol.getVertex(v);
for (int nei = vertex.neiBegin(); nei != vertex.neiEnd(); nei = vertex.neiNext(nei))
{
int edge = vertex.neiEdge(nei);
QueryMolecule::Bond &bond = mol.getBond(edge);
// DP TODO: implement smth. like Node::possibleOtherValueExcept() ...
bool can_be_double = bond.possibleValue(QueryMolecule::BOND_ORDER, BOND_DOUBLE);
bool can_be_triple = bond.possibleValue(QueryMolecule::BOND_ORDER, BOND_TRIPLE);
bool can_be_arom = bond.possibleValue(QueryMolecule::BOND_ORDER, BOND_AROMATIC);
if (can_be_double || can_be_triple || can_be_arom)
{
QS_DEF(Array<int>, sites);
mol.getAllowedRGroups(v, sites);
for (int j = 0; j < sites.size(); j++)
{
rgroups_attached_single.expandFill(sites[j] + 1, true);
rgroups_attached_single[sites[j]] = false;
}
}
}
}
}
rgroups_attached_single.expandFill(n_rgroups + 1, true);
for (int i = 1; i <= n_rgroups; i++)
{
PtrPool<BaseMolecule> &frags = rgroups.getRGroup(i).fragments;
for (int j = frags.begin(); j != frags.end(); j = frags.next(j))
{
QueryMolecule &fragment = frags[j]->asQueryMolecule();
aromatized |= _aromatizeRGroupFragment(fragment, rgroups_attached_single[i], options);
}
}
return aromatized;
}
bool MoleculeSubstructureMatcher::matchQueryAtom
(QueryMolecule::Atom *query, BaseMolecule &target, int super_idx,
FragmentMatchCache *fmcache, dword flags)
{
int i;
switch (query->type)
{
case QueryMolecule::OP_NONE:
return true;
case QueryMolecule::OP_AND:
for (i = 0; i < query->children.size(); i++)
if (!matchQueryAtom(query->child(i), target, super_idx, fmcache, flags))
return false;
return true;
case QueryMolecule::OP_OR:
for (i = 0; i < query->children.size(); i++)
if (matchQueryAtom(query->child(i), target,
super_idx, fmcache, flags))
return true;
return false;
case QueryMolecule::OP_NOT:
return !matchQueryAtom(query->child(0), target, super_idx, fmcache,
flags ^ MATCH_DISABLED_AS_TRUE);
case QueryMolecule::ATOM_NUMBER:
return query->valueWithinRange(target.getAtomNumber(super_idx));
case QueryMolecule::ATOM_PSEUDO:
return target.isPseudoAtom(super_idx) &&
strcmp(query->alias.ptr(), target.getPseudoAtom(super_idx)) == 0;
case QueryMolecule::ATOM_RSITE:
return true;
case QueryMolecule::ATOM_ISOTOPE:
return query->valueWithinRange(target.getAtomIsotope(super_idx));
case QueryMolecule::ATOM_CHARGE:
{
if (flags & MATCH_ATOM_CHARGE)
return query->valueWithinRange(target.getAtomCharge(super_idx));
return (flags & MATCH_DISABLED_AS_TRUE) != 0;
}
case QueryMolecule::ATOM_RADICAL:
{
if (target.isPseudoAtom(super_idx) || target.isRSite(super_idx))
return false;
return query->valueWithinRange(target.getAtomRadical(super_idx));
}
case QueryMolecule::ATOM_VALENCE:
{
if (flags & MATCH_ATOM_VALENCE)
{
if (target.isPseudoAtom(super_idx) || target.isRSite(super_idx))
return false;
return query->valueWithinRange(target.getAtomValence(super_idx));
}
return (flags & MATCH_DISABLED_AS_TRUE) != 0;
}
case QueryMolecule::ATOM_CONNECTIVITY:
{
int conn = target.getVertex(super_idx).degree();
if (!target.isPseudoAtom(super_idx) && !target.isRSite(super_idx))
conn += target.asMolecule().getImplicitH(super_idx);
return query->valueWithinRange(conn);
}
case QueryMolecule::ATOM_TOTAL_BOND_ORDER:
{
// TODO: target.isPseudoAtom(super_idx) || target.isRSite(super_idx)
return query->valueWithinRange(target.asMolecule().getAtomConnectivity(super_idx));
}
case QueryMolecule::ATOM_TOTAL_H:
{
if (target.isPseudoAtom(super_idx) || target.isRSite(super_idx))
return false;
return query->valueWithinRange(target.getAtomTotalH(super_idx));
}
case QueryMolecule::ATOM_SUBSTITUENTS:
return query->valueWithinRange(target.getAtomSubstCount(super_idx));
case QueryMolecule::ATOM_SSSR_RINGS:
return query->valueWithinRange(target.vertexCountSSSR(super_idx));
case QueryMolecule::ATOM_SMALLEST_RING_SIZE:
return query->valueWithinRange(target.vertexSmallestRingSize(super_idx));
case QueryMolecule::ATOM_RING_BONDS:
case QueryMolecule::ATOM_RING_BONDS_AS_DRAWN:
return query->valueWithinRange(target.getAtomRingBondsCount(super_idx));
case QueryMolecule::ATOM_UNSATURATION:
return !target.isSaturatedAtom(super_idx);
case QueryMolecule::ATOM_FRAGMENT:
{
if (fmcache == 0)
throw Error("unexpected 'fragment' constraint");
QueryMolecule *fragment = query->fragment.get();
const char *smarts = fragment->fragment_smarts.ptr();
if (fragment->vertexCount() == 0)
throw Error("empty fragment");
if (smarts != 0 && strlen(smarts) > 0)
{
fmcache->expand(super_idx + 1);
int *value = fmcache->at(super_idx).at2(smarts);
if (value != 0)
return *value != 0;
}
MoleculeSubstructureMatcher matcher(target.asMolecule());
matcher.not_ignore_first_atom = true;
matcher.setQuery(*fragment);
matcher.fmcache = fmcache;
bool result = matcher.fix(fragment->vertexBegin(), super_idx);
if (result)
result = matcher.find();
if (smarts != 0 && strlen(smarts) > 0)
{
fmcache->expand(super_idx + 1);
fmcache->at(super_idx).insert(smarts, result ? 1 : 0);
}
return result;
}
case QueryMolecule::ATOM_AROMATICITY:
return query->valueWithinRange(target.getAtomAromaticity(super_idx));
case QueryMolecule::HIGHLIGHTING:
return query->valueWithinRange((int)target.isAtomHighlighted(super_idx));
default:
throw Error("bad query atom type: %d", query->type);
}
}
bool MoleculeSubstructureMatcher::_shouldUnfoldTargetHydrogens (QueryMolecule &query, bool is_fragment, bool disable_folding_query_h)
{
int i, j;
for (i = query.vertexBegin(); i != query.vertexEnd(); i = query.vertexNext(i))
{
// skip R-atoms
if (query.isRSite(i))
continue;
if (query.possibleAtomNumberAndIsotope(i, ELEM_H, 0))
{
const Vertex &vertex = query.getVertex(i);
// Degree 2 or higher => definilely not a hydrogen
if (vertex.degree() > 1)
continue;
// Can be lone hydrogen?
if (vertex.degree() == 0)
return true;
// degree is 1 at this point
int edge_idx = vertex.neiEdge(vertex.neiBegin());
// is it is double or triple bond => not hydrogen
if (query.getBondOrder(edge_idx) > 1)
continue;
// ring bond?
if (query.getBondTopology(edge_idx) == TOPOLOGY_RING)
continue;
// can be something other than hydrogen?
if (query.getAtomNumber(i) == -1)
return true;
if (is_fragment && i == query.vertexBegin())
// If first atom in a fragment is hydrogen then hydrogens should
// be unfolded because of the matching logic: when fragment will be
// matched this first hydrogen should match some atom.
// If hydrogens is not be unfolded in this case then
// [$([#1][N])]C will not match NC.
return true;
// If we need to find all embeddings then query hydrogens cannot be ignored:
// For example, if we are searching number of matcher for N-[#1] in N then
// it should 3 instead of 1
if (disable_folding_query_h)
return true;
// Check if hydrogen forms a cis-trans bond or stereocenter
int nei_vertex_idx = vertex.neiVertex(vertex.neiBegin());
if (query.stereocenters.exists(nei_vertex_idx))
return true;
// For example for this query hydrogens should be unfolded: [H]\\C=C/C
const Vertex &nei_vertex = query.getVertex(nei_vertex_idx);
for (int nei = nei_vertex.neiBegin(); nei != nei_vertex.neiEnd(); nei = nei_vertex.neiNext(nei))
{
int edge = nei_vertex.neiEdge(nei);
if (query.cis_trans.getParity(edge) != 0)
return true;
}
}
if (_shouldUnfoldTargetHydrogens_A(&query.getAtom(i), is_fragment, disable_folding_query_h))
return true;
}
MoleculeRGroups &rgroups = query.rgroups;
int n_rgroups = rgroups.getRGroupCount();
for (i = 1; i <= n_rgroups; i++)
{
PtrPool<BaseMolecule> &frags = rgroups.getRGroup(i).fragments;
for (j = frags.begin(); j != frags.end(); j = frags.next(j))
if (_shouldUnfoldTargetHydrogens(frags[j]->asQueryMolecule(), is_fragment, disable_folding_query_h))
return true;
}
return false;
}