void IndigoInchi::saveMoleculeIntoInchi (Molecule &mol, Array<char> &inchi)
{
inchi_Input input;
QS_DEF(Array<inchi_Atom>, atoms);
QS_DEF(Array<inchi_Stereo0D>, stereo);
// Check if structure has aromatic bonds
bool has_aromatic = false;
for (int e = mol.edgeBegin(); e != mol.edgeEnd(); e = mol.edgeNext(e))
if (mol.getBondOrder(e) == BOND_AROMATIC)
{
has_aromatic = true;
break;
}
Molecule *target = &mol;
Obj<Molecule> dearom;
if (has_aromatic)
{
dearom.create();
dearom->clone(mol, 0, 0);
try
{
dearom->dearomatize();
}
catch (DearomatizationsGroups::Error &)
{
}
target = dearom.get();
}
generateInchiInput(*target, input, atoms, stereo);
inchi_Output output;
int ret = GetINCHI(&input, &output);
if (output.szMessage)
warning.readString(output.szMessage, true);
if (output.szLog)
log.readString(output.szLog, true);
if (output.szAuxInfo)
auxInfo.readString(output.szAuxInfo, true);
if (ret != inchi_Ret_OKAY && ret != inchi_Ret_WARNING)
{
// Construct error before dispoing inchi output to preserve error message
IndigoError error("Indigo-InChI: InChI generation failed: %s. Code: %d.", output.szMessage, ret);
FreeINCHI(&output);
throw error;
}
inchi.readString(output.szInChI, true);
FreeINCHI(&output);
}
bool AromaticityMatcher::match (int *core_sub, int *core_super)
{
// Check if detailed checking is necessary
bool needCheck = false;
for (int i = _query.edgeBegin(); i != _query.edgeEnd(); i = _query.edgeNext(i))
{
if (!_query.getBond(i).hasConstraint(QueryMolecule::BOND_ORDER))
continue;
if (_matching_edges_state[i] == AROMATIC && _query.getBondOrder(i) != BOND_AROMATIC)
{
needCheck = true;
break;
}
}
if (!needCheck)
return true;
// By our rules submolecule in the query, that maps on aromatic bonds in
// the target, must have aromatic bond configuration to match the target.
// To check this such submolecule from query molecule is extracted, then
// all bonds are marked as aromatic, and then dearomatizer tries to find
// aromatic bonds configuration with partially fixed bonds.
// 1. Extract query submolecule that maps on aromatic bonds. It is the same as in target.
// Set skip all additional informatio during copying
QS_DEF(Array<int>, mapping);
mapping.clear();
for (int v_idx = _query.vertexBegin();
v_idx < _query.vertexEnd();
v_idx = _query.vertexNext(v_idx))
{
int target_idx = core_sub[v_idx];
if (target_idx < 0)
continue;
mapping.push(target_idx);
}
QS_DEF(Array<int>, edges);
QS_DEF(Array<int>, base_edges_mask);
edges.clear();
base_edges_mask.clear_resize(_base.edgeEnd());
base_edges_mask.zerofill();
for (int e_idx = _query.edgeBegin();
e_idx < _query.edgeEnd();
e_idx = _query.edgeNext(e_idx))
{
const Edge &e = _query.getEdge(e_idx);
if (core_sub[e.beg] < 0 || core_sub[e.end] < 0)
continue;
int target_idx = _base.findEdgeIndex(core_sub[e.beg], core_sub[e.end]);
if (target_idx == -1)
throw Error("(AromaticityMatcher::match) target edge wasn't found");
edges.push(target_idx);
base_edges_mask[target_idx] = 1;
}
QS_DEF(Array<int>, inv_mapping);
_submolecule->makeEdgeSubmolecule(_base, mapping, edges, &inv_mapping, SKIP_ALL);
QS_DEF(Array<int>, external_conn);
external_conn.resize(_submolecule->vertexEnd());
external_conn.zerofill();
// Calculate external connectivity
for (int i = 0; i < mapping.size(); i++)
{
int base_idx = mapping[i];
const Vertex &v = _base.getVertex(base_idx);
int cur_external_conn = 0;
for (int ni = v.neiBegin(); ni != v.neiEnd(); ni = v.neiNext(ni))
{
int ni_edge = v.neiEdge(ni);
if (!base_edges_mask[ni_edge])
{
int bond_order_diff = _base.getBondOrder(ni_edge);
if (bond_order_diff == BOND_AROMATIC)
bond_order_diff = 1;
cur_external_conn += bond_order_diff;
}
}
external_conn[i] = cur_external_conn;
}
// 1b. Find bonds in aromatic rings in query and skip aromatic
// bonds that are not in cycles
QS_DEF(Array<int>, is_edge_in_aromatic_cycle);
is_edge_in_aromatic_cycle.clear_resize(_submolecule->edgeEnd());
is_edge_in_aromatic_cycle.zerofill();
// At first just mark aromatic bonds
for (int e_idx = _submolecule->edgeBegin();
e_idx < _submolecule->edgeEnd();
e_idx = _submolecule->edgeNext(e_idx))
{
if (_submolecule->getBondOrder(e_idx) == BOND_AROMATIC)
is_edge_in_aromatic_cycle[e_idx] = 1;
}
Filter aromatic_edge_filter(is_edge_in_aromatic_cycle.ptr(), Filter::EQ, 1);
SpanningTree arom_edges_st(_submolecule.ref(), 0, &aromatic_edge_filter);
// Store in is_edge_in_aromatic_cycle marks about such bonds
is_edge_in_aromatic_cycle.zerofill();
arom_edges_st.markAllEdgesInCycles(is_edge_in_aromatic_cycle.ptr(), 1);
enum { AROMATIC_BOND_NOT_IN_AROMATIC_CYCLE = 2 };
for (int e_idx = _submolecule->edgeBegin();
e_idx < _submolecule->edgeEnd();
e_idx = _submolecule->edgeNext(e_idx))
{
if (_submolecule->getBondOrder(e_idx) == BOND_AROMATIC &&
is_edge_in_aromatic_cycle[e_idx] == 0)
{
// Such bond is marked as aromatic but it isn't in aromatic ring
// Here just change bond order to nonaromatic (single) and later
// check if such query bond can be aromatic (not by aromatizer)
if (_submolecule->isQueryMolecule())
{
QueryMolecule &qmol = _submolecule->asQueryMolecule();
AutoPtr<QueryMolecule::Bond> bond(qmol.releaseBond(e_idx));
bond->removeConstraints(QueryMolecule::BOND_ORDER);
AutoPtr<QueryMolecule::Bond> arom_bond(
new QueryMolecule::Bond(QueryMolecule::BOND_ORDER, BOND_AROMATIC));
qmol.resetBond(e_idx, QueryMolecule::Bond::und(bond.release(), arom_bond.release()));
}
else
_submolecule->asMolecule().setBondOrder(e_idx, BOND_SINGLE, false);
is_edge_in_aromatic_cycle[e_idx] = AROMATIC_BOND_NOT_IN_AROMATIC_CYCLE;
}
}
// 2. Try to find suitable dearomatization
QS_DEF(DearomatizationsStorage, dearomatizations);
// Dearomatizer and DearomatizationMatcher will be created on demand
Obj<Dearomatizer> dearomatizer;
Obj<DearomatizationMatcher> dearomatizationMatcher;
// Check edges
for (int e = _submolecule->edgeBegin(); e != _submolecule->edgeEnd(); e = _submolecule->edgeNext(e))
{
const Edge &edge = _submolecule->getEdge(e);
int target_edge_index =
_base.findEdgeIndex(mapping[edge.beg], mapping[edge.end]);
const Edge &target_edge = _base.getEdge(target_edge_index);
int query_edge_index =
_query.findEdgeIndex(core_super[target_edge.beg], core_super[target_edge.end]);
if (query_edge_index == -1)
throw Error("(AromaticityMatcher::match) query edge wasn't found");
if (_matching_edges_state[query_edge_index] != AROMATIC)
// Such target bond isn't aromatic. So we don't need to fix such bond
continue;
QueryMolecule::Bond &qbond = _query.getBond(query_edge_index);
bool can_have_arom_order = qbond.possibleValuePair(
QueryMolecule::BOND_ORDER, BOND_AROMATIC,
QueryMolecule::BOND_TOPOLOGY, TOPOLOGY_RING);
if (can_have_arom_order)
// This query bond is aromatic.
continue;
if (is_edge_in_aromatic_cycle[e] == AROMATIC_BOND_NOT_IN_AROMATIC_CYCLE)
{
// Such bond cannot be aromatic without aromatic
// cycle because 'can_have_arom_order' is false
return false;
}
bool has_other = !qbond.hasNoConstraintExcept(QueryMolecule::BOND_ORDER,
QueryMolecule::BOND_TOPOLOGY);
if (has_other)
throw Error("Only bond with order and topology constraints are supported");
bool can_have_single = qbond.possibleValuePair(
QueryMolecule::BOND_ORDER, BOND_SINGLE,
QueryMolecule::BOND_TOPOLOGY, TOPOLOGY_RING);
bool can_have_double = qbond.possibleValuePair(
QueryMolecule::BOND_ORDER, BOND_DOUBLE,
QueryMolecule::BOND_TOPOLOGY, TOPOLOGY_RING);
// Check if query bond can be only single or only double
if (can_have_single && can_have_double)
// Don't fix such bond. It can be single/double, single/aromatic and etc.
continue;
if (!can_have_single && !can_have_double)
// Bond without and specification and etc.
continue;
// Find dearomatization
if (dearomatizer.get() == NULL)
{
dearomatizer.create(_submolecule.ref(), external_conn.ptr(), _arom_options);
dearomatizer->enumerateDearomatizations(dearomatizations);
dearomatizationMatcher.create(dearomatizations,
_submolecule.ref(), external_conn.ptr());
}
// Fix bond
int type = can_have_single ? BOND_SINGLE : BOND_DOUBLE;
if (!dearomatizationMatcher->isAbleToFixBond(e, type))
return false;
dearomatizationMatcher->fixBond(e, type);
}
return true;
}
void CrfSaver::_writeMolecule (Molecule &molecule)
{
Obj<CmfSaver> saver;
int i;
if (_encoder.get() != 0)
saver.create(_encoder.ref());
else
saver.create(_output);
QS_DEF(Array<int>, atom_flags);
QS_DEF(Array<int>, bond_flags);
if (_atom_stereo_flags != 0)
{
atom_flags.clear_resize(molecule.vertexEnd());
atom_flags.zerofill();
for (i = molecule.vertexBegin(); i != molecule.vertexEnd(); i = molecule.vertexNext(i))
if (_atom_stereo_flags[i] == STEREO_RETAINS)
atom_flags[i] = 1;
else if (_atom_stereo_flags[i] == STEREO_INVERTS)
atom_flags[i] = 2;
saver->atom_flags = atom_flags.ptr();
}
if (_bond_rc_flags != 0)
{
bond_flags.clear_resize(molecule.edgeEnd());
bond_flags.zerofill();
for (i = molecule.edgeBegin(); i != molecule.edgeEnd(); i = molecule.edgeNext(i))
{
if (_bond_rc_flags[i] & RC_UNCHANGED)
bond_flags[i] |= 1;
if (_bond_rc_flags[i] & RC_MADE_OR_BROKEN)
bond_flags[i] |= 2;
if (_bond_rc_flags[i] & RC_ORDER_CHANGED)
bond_flags[i] |= 4;
}
saver->bond_flags = bond_flags.ptr();
}
saver->save_bond_dirs = save_bond_dirs;
saver->save_highlighting = save_highlighting;
saver->save_mapping = save_mapping;
saver->saveMolecule(molecule);
if (_aam != 0)
_writeAam(_aam, saver->getAtomSequence());
if (xyz_output != 0)
{
if (xyz_output == &_output && _encoder.get() != 0)
_encoder->finish();
saver->saveXyz(*xyz_output);
if (xyz_output == &_output && _encoder.get() != 0)
_encoder->start();
}
}
void CrfLoader::_loadMolecule (Molecule &molecule)
{
Obj<CmfLoader> loader;
int i;
if (_decoder.get() != 0)
loader.create(_decoder.ref());
else
loader.create(_scanner);
QS_DEF(Array<int>, atom_flags);
QS_DEF(Array<int>, bond_flags);
loader->atom_flags = &atom_flags;
loader->bond_flags = &bond_flags;
loader->version = version;
loader->loadMolecule(molecule);
bool has_mapping = loader->has_mapping;
if (_atom_stereo_flags != 0)
{
_atom_stereo_flags->clear_resize(molecule.vertexCount());
_atom_stereo_flags->zerofill();
for (i = 0; i < molecule.vertexCount(); i++)
{
int idx = i;
if (has_mapping)
idx = loader->inv_atom_mapping_to_restore[i];
if (atom_flags[i] & 1)
_atom_stereo_flags->at(idx) |= STEREO_RETAINS;
if (atom_flags[i] & 2)
_atom_stereo_flags->at(idx) |= STEREO_INVERTS;
}
}
if (_bond_rc_flags != 0)
{
_bond_rc_flags->clear_resize(molecule.edgeCount());
_bond_rc_flags->zerofill();
for (i = 0; i < molecule.edgeCount(); i++)
{
int idx = i;
if (has_mapping)
idx = loader->inv_bond_mapping_to_restore[i];
if (bond_flags[i] & 1)
_bond_rc_flags->at(idx) |= RC_UNCHANGED;
if (bond_flags[i] & 2)
_bond_rc_flags->at(idx) |= RC_MADE_OR_BROKEN;
if (bond_flags[i] & 4)
_bond_rc_flags->at(idx) |= RC_ORDER_CHANGED;
}
}
if (_aam != 0)
{
_aam->clear_resize(molecule.vertexCount());
_aam->zerofill();
for (i = 0; i < molecule.vertexCount(); i++)
{
int value;
if (_decoder.get() != 0)
value = _decoder->get();
else
value = _scanner.readByte();
int idx = i;
if (has_mapping)
idx = loader->inv_atom_mapping_to_restore[i];
_aam->at(idx) = value - 1;
}
}
if (xyz_scanner != 0)
loader->loadXyz(*xyz_scanner);
}
