/// <summary>
/// Convert a CDK <see cref="IBond"/> to a Beam edge.
/// </summary>
/// <param name="b">the CDK bond</param>
/// <param name="indices">map of atom indices</param>
/// <returns>a Beam edge</returns>
/// <exception cref="ArgumentException">the bond did not have 2 atoms or an unsupported order</exception>
/// <exception cref="NullReferenceException">the bond order was undefined</exception>
internal static Edge ToBeamEdge(IBond b, SmiFlavors flavour, Dictionary <IAtom, int> indices)
{
CheckArgument(b.Atoms.Count == 2, "Invalid number of atoms on bond");
var u = indices[b.Begin];
var v = indices[b.End];
return(ToBeamEdgeLabel(b, flavour).CreateEdge(u, v));
}
static Beam.Graph Convert(IAtomContainer ac, bool perceiveAromaticity, SmiFlavors options)
{
AtomContainerManipulator.PercieveAtomTypesAndConfigureAtoms(ac);
CDK.HydrogenAdder.AddImplicitHydrogens(ac);
if (perceiveAromaticity)
{
Aromaticity.CDKLegacy.Apply(ac);
}
return(new CDKToBeam(options).ToBeamGraph(ac));
}
/// <summary>
/// Convert an CDK <see cref="IAtom"/> to a Beam Atom. The symbol and implicit
/// hydrogen count are not optional. If the symbol is not supported by the
/// SMILES notation (e.g. 'R1') the element will automatically default to
/// Unknown ('*').
/// </summary>
/// <param name="a">cdk Atom instance</param>
/// <returns>a Beam atom</returns>
/// <exception cref="NullReferenceException">the atom had an undefined symbol or implicit hydrogen count</exception>
static Beam.IAtom ToBeamAtom(IAtom a, SmiFlavors flavour)
{
var aromatic = SmiFlavorTool.IsSet(flavour, SmiFlavors.UseAromaticSymbols) && a.IsAromatic;
var charge = a.FormalCharge;
string symbol = CheckNotNull(a.Symbol, "An atom had an undefined symbol");
var element = Beam.Element.OfSymbol(symbol);
if (element == null)
{
element = Beam.Element.Unknown;
}
var ab = aromatic ? AtomBuilder.Aromatic(element) : AtomBuilder.Aliphatic(element);
// CDK leaves nulls on pseudo atoms - we need to check this special case
var hCount = a.ImplicitHydrogenCount;
if (element == Beam.Element.Unknown)
{
ab.NumOfHydrogens(hCount ?? 0);
}
else
{
ab.NumOfHydrogens(CheckNotNull(hCount, "One or more atoms had an undefined number of implicit hydrogens"));
}
if (charge.HasValue)
{
ab.Charge(charge.Value);
}
// use the mass number to specify isotope?
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.AtomicMass | SmiFlavors.AtomicMassStrict))
{
var massNumber = a.MassNumber;
if (massNumber != null)
{
ab.Isotope(massNumber.Value);
}
}
var atomClass = a.GetProperty <int?>(CDKPropertyName.AtomAtomMapping);
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.AtomAtomMap) && atomClass != null)
{
ab.AtomClass(atomClass.Value);
}
return(ab.Build());
}
/// <summary>
/// Given a molecule (possibly disconnected) compute the labels which
/// would order the atoms by increasing canonical labelling. If the SMILES
/// are isomeric (i.e. stereo and isotope specific) the InChI numbers are
/// used. These numbers are loaded via reflection and the 'cdk-inchi' module
/// should be present on the path.
/// </summary>
/// <param name="molecule">the molecule to</param>
/// <returns>the permutation</returns>
/// <seealso cref="Canon"/>
private static int[] Labels(SmiFlavors flavour, IAtomContainer molecule)
{
// FIXME: use SmiOpt.InChiLabelling
var labels = SmiFlavorTool.IsSet(flavour, SmiFlavors.Isomeric)
? InChINumbers(molecule)
: Canon.Label(molecule, GraphUtil.ToAdjList(molecule), CreateComparator(molecule, flavour));
var cpy = new int[labels.Length];
for (int i = 0; i < labels.Length; i++)
{
cpy[i] = (int)labels[i] - 1;
}
return(cpy);
}
/// <summary>
/// Convert a CDK <see cref="IAtomContainer"/> to a Beam ChemicalGraph. The graph
/// can when be written directly as to a SMILES or manipulated further (e.g
/// canonical ordering/standard-form and other normalisations).
/// </summary>
/// <param name="ac">an atom container instance</param>
/// <returns>the Beam ChemicalGraph for additional manipulation</returns>
internal static Graph ToBeamGraph(IAtomContainer ac, SmiFlavors flavour)
{
int order = ac.Atoms.Count;
var gb = GraphBuilder.Create(order);
var indices = new Dictionary <IAtom, int>(order);
foreach (var a in ac.Atoms)
{
indices[a] = indices.Count;
gb.Add(ToBeamAtom(a, flavour));
}
foreach (var b in ac.Bonds)
{
gb.Add(ToBeamEdge(b, flavour, indices));
}
// configure stereo-chemistry by encoding the stereo-elements
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.Stereo))
{
foreach (var se in ac.StereoElements)
{
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.StereoTetrahedral) &&
se is ITetrahedralChirality)
{
AddTetrahedralConfiguration((ITetrahedralChirality)se, gb, indices);
}
else if (SmiFlavorTool.IsSet(flavour, SmiFlavors.StereoCisTrans) &&
se is IDoubleBondStereochemistry)
{
AddGeometricConfiguration((IDoubleBondStereochemistry)se, flavour, gb, indices);
}
else if (SmiFlavorTool.IsSet(flavour, SmiFlavors.StereoExTetrahedral) &&
se is ExtendedTetrahedral)
{
AddExtendedTetrahedralConfiguration((ExtendedTetrahedral)se, gb, indices);
}
else if (SmiFlavorTool.IsSet(flavour, SmiFlavors.StereoExCisTrans) &&
se is ExtendedCisTrans)
{
AddExtendedCisTransConfig((ExtendedCisTrans)se, gb, indices, ac);
}
}
}
return(gb.Build());
}
/// <summary>
/// Convert a CDK <see cref="IBond"/> to the Beam edge label type.
/// </summary>
/// <param name="b">cdk bond</param>
/// <returns>the edge label for the Beam edge</returns>
/// <exception cref="NullReferenceException">the bond order was null and the bond was not-aromatic</exception>
/// <exception cref="ArgumentException">the bond order could not be converted</exception>
private static Bond ToBeamEdgeLabel(IBond b, SmiFlavors flavour)
{
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.UseAromaticSymbols) && b.IsAromatic)
{
if (!b.Begin.IsAromatic || !b.End.IsAromatic)
{
throw new InvalidOperationException("Aromatic bond connects non-aromatic atomic atoms");
}
return(Bond.Aromatic);
}
if (b.Order.IsUnset())
{
throw new CDKException("A bond had undefined order, possible query bond?");
}
var order = b.Order;
switch (order)
{
case BondOrder.Single:
return(Bond.Single);
case BondOrder.Double:
return(Bond.Double);
case BondOrder.Triple:
return(Bond.Triple);
case BondOrder.Quadruple:
return(Bond.Quadruple);
default:
throw new CDKException("Unsupported bond order: " + order);
}
}
static Beam.Graph Convert(IAtomContainer ac, SmiFlavors options)
{
return(Convert(ac, false, options));
}
internal static string Generate(CxSmilesState state, SmiFlavors opts, int[] components, int[] ordering)
{
if (!SmiFlavorTool.IsSet(opts, SmiFlavors.CxSmilesWithCoords))
{
return("");
}
var invorder = Inverse(ordering);
var sb = new StringBuilder();
sb.Append(' ');
sb.Append('|');
//int invComp(int a, int b) => invorder[a].CompareTo(invorder[b]);
int comp(int a, int b) => ordering[a].CompareTo(ordering[b]);
// Fragment Grouping
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxFragmentGroup) &&
state.fragGroups != null &&
state.fragGroups.Any())
{
int maxCompId = 0;
foreach (int compId_ in components)
{
if (compId_ > maxCompId)
{
maxCompId = compId_;
}
}
// get the output component order
var compMap = new int[maxCompId + 1];
int compId = 1;
foreach (int idx in invorder)
{
var component = components[idx];
if (compMap[component] == 0)
{
compMap[component] = compId++;
}
}
// index vs number, we need to output index
for (int i = 0; i < compMap.Length; i++)
{
compMap[i]--;
}
int compComp(int a, int b) => compMap[a].CompareTo(compMap[b]);
var fragGroupCpy = new List <List <int> >(state.fragGroups);
foreach (var idxs in fragGroupCpy)
{
idxs.Sort(compComp);
}
fragGroupCpy.Sort((a, b) => CxSmilesGenerator.Compare(compComp, a, b));
// C1=CC=CC=C1.C1=CC=CC=C1.[OH-].[Na+]>> |f:0.1,2.3,c:0,2,4,6,8,10|
sb.Append('f');
sb.Append(':');
for (int i = 0; i < fragGroupCpy.Count; i++)
{
if (i > 0)
{
sb.Append(',');
}
AppendIntegers(compMap, '.', sb, fragGroupCpy[i]);
}
}
// Atom Labels
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxAtomLabel) &&
state.atomLabels != null && state.atomLabels.Any())
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append('$');
int nonempty_cnt = 0;
foreach (int idx in invorder)
{
if (!state.atomLabels.TryGetValue(idx, out string label))
{
label = "";
}
else
{
nonempty_cnt++;
}
sb.Append(Encode_alias(label));
// don't need to write anymore more ';'
if (nonempty_cnt == state.atomLabels.Count)
{
break;
}
sb.Append(";");
}
sb.Append('$');
}
// Atom Values
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxAtomValue) &&
state.atomValues != null && state.atomValues.Any())
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append("$_AV:");
int nonempty_cnt = 0;
foreach (int idx in invorder)
{
var label = state.atomValues[idx];
if (string.IsNullOrEmpty(label))
{
label = "";
}
else
{
nonempty_cnt++;
}
sb.Append(Encode_alias(label));
// don't need to write anymore more ';'
if (nonempty_cnt == state.atomValues.Count)
{
break;
}
sb.Append(";");
}
sb.Append('$');
}
// 2D/3D Coordinates
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxCoordinates) &&
state.atomCoords != null && state.atomCoords.Any())
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append('(');
for (int i = 0; i < ordering.Length; i++)
{
var xyz = state.atomCoords[invorder[i]];
if (i != 0)
{
sb.Append(';');
}
if (xyz[0] != 0)
{
sb.Append(Strings.ToSimpleString(xyz[0], 2));
}
sb.Append(',');
if (xyz[1] != 0)
{
sb.Append(Strings.ToSimpleString(xyz[1], 2));
}
sb.Append(',');
if (xyz[2] != 0)
{
sb.Append(Strings.ToSimpleString(xyz[2], 2));
}
}
sb.Append(')');
}
// Multi-center/Positional variation bonds
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxMulticenter) &&
state.positionVar != null &&
state.positionVar.Any())
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append('m');
sb.Append(':');
var multicenters = new List <KeyValuePair <int, IList <int> > >(state.positionVar);
// consistent output order
multicenters.Sort((a, b) => comp(a.Key, b.Key));
for (int i = 0; i < multicenters.Count; i++)
{
if (i != 0)
{
sb.Append(',');
}
var e = multicenters[i];
sb.Append(ordering[e.Key]);
sb.Append(':');
var vals = new List <int>(e.Value);
vals.Sort(comp);
AppendIntegers(ordering, '.', sb, vals);
}
}
// *CCO* |$_AP1;;;;_AP2$,Sg:n:1,2,3::ht|
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxPolymer) &&
state.sgroups != null && state.sgroups.Any())
{
var sgroups = new List <PolymerSgroup>(state.sgroups);
foreach (PolymerSgroup psgroup in sgroups)
{
psgroup.atomset.Sort(comp);
}
sgroups.Sort((a, b) =>
{
int cmp = 0;
cmp = string.CompareOrdinal(a.type, b.type);
if (cmp != 0)
{
return(cmp);
}
cmp = CxSmilesGenerator.Compare(comp, a.atomset, b.atomset);
return(cmp);
});
for (int i = 0; i < sgroups.Count; i++)
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append("Sg:");
var sgroup = sgroups[i];
sb.Append(sgroup.type);
sb.Append(':');
AppendIntegers(ordering, ',', sb, sgroup.atomset);
sb.Append(':');
if (sgroup.subscript != null)
{
sb.Append(sgroup.subscript);
}
sb.Append(':');
if (sgroup.supscript != null)
{
sb.Append(sgroup.supscript.ToLowerInvariant());
}
}
}
// [C]1[CH][CH]CCC1 |^1:1,2,^3:0|
if (SmiFlavorTool.IsSet(opts, SmiFlavors.CxRadical) &&
state.atomRads != null &&
state.atomRads.Any())
{
var radinv = new SortedDictionary <CxSmilesState.Radical, List <int> >();
foreach (var e in state.atomRads)
{
if (!radinv.TryGetValue(e.Value, out List <int> idxs))
{
radinv[e.Value] = idxs = new List <int>();
}
idxs.Add(e.Key);
}
foreach (var e in radinv)
{
if (sb.Length > 2)
{
sb.Append(',');
}
sb.Append('^');
sb.Append((int)e.Key + 1);
sb.Append(':');
e.Value.Sort(comp);
AppendIntegers(ordering, ',', sb, e.Value);
}
}
sb.Append('|');
if (sb.Length <= 3)
{
return("");
}
else
{
return(sb.ToString());
}
}
internal CDKToBeam(SmiFlavors flavour)
{
this.flavour = flavour;
}
/// <summary>
/// Add double-bond stereo configuration to the Beam GraphBuilder.
/// </summary>
/// <param name="dbs">stereo element specifying double-bond configuration</param>
/// <param name="gb">the current graph builder</param>
/// <param name="indices">atom indices</param>
private static void AddGeometricConfiguration(IDoubleBondStereochemistry dbs, SmiFlavors flavour, GraphBuilder gb, Dictionary <IAtom, int> indices)
{
var db = dbs.StereoBond;
var bs = dbs.Bonds;
// don't try to set a configuration on aromatic bonds
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.UseAromaticSymbols) && db.IsAromatic)
{
return;
}
var u = indices[db.Begin];
var v = indices[db.End];
// is bs[0] always connected to db.Atom(0)?
var x = indices[bs[0].GetOther(db.Begin)];
var y = indices[bs[1].GetOther(db.End)];
if (dbs.Stereo == DoubleBondConformation.Together)
{
gb.Geometric(u, v).Together(x, y);
}
else
{
gb.Geometric(u, v).Opposite(x, y);
}
}
internal static bool IsSet(this SmiFlavors opts, SmiFlavors opt)
{
return((opts & opt) != 0);
}
static IComparer <IAtom> CreateComparator(IAtomContainer mol, SmiFlavors flavor)
{
return(new Comparer(mol, flavor));
}
public Comparer(IAtomContainer mol, SmiFlavors flavor)
{
this.mol = mol;
this.flavor = flavor;
}
// Creates a CxSmilesState from a molecule with atom labels, repeat units, multi-center bonds etc
private static CxSmilesState GetCxSmilesState(SmiFlavors flavour, IAtomContainer mol)
{
CxSmilesState state = new CxSmilesState
{
atomCoords = new List <double[]>(),
coordFlag = false
};
// set the atom labels, values, and coordinates,
// and build the atom->idx map required by other parts
var atomidx = new Dictionary <IAtom, int>();
for (int idx = 0; idx < mol.Atoms.Count; idx++)
{
IAtom atom = mol.Atoms[idx];
if (atom is IPseudoAtom)
{
if (state.atomLabels == null)
{
state.atomLabels = new SortedDictionary <int, string>();
}
IPseudoAtom pseudo = (IPseudoAtom)atom;
if (pseudo.AttachPointNum > 0)
{
state.atomLabels[idx] = "_AP" + pseudo.AttachPointNum;
}
else
{
if (!"*".Equals(pseudo.Label, StringComparison.Ordinal))
{
state.atomLabels[idx] = pseudo.Label;
}
}
}
object comment = atom.GetProperty <object>(CDKPropertyName.Comment);
if (comment != null)
{
if (state.atomValues == null)
{
state.atomValues = new SortedDictionary <int, string>();
}
state.atomValues[idx] = comment.ToString();
}
atomidx[atom] = idx;
var p2 = atom.Point2D;
var p3 = atom.Point3D;
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.Cx2dCoordinates) && p2 != null)
{
state.atomCoords.Add(new double[] { p2.Value.X, p2.Value.Y, 0 });
state.coordFlag = true;
}
else if (SmiFlavorTool.IsSet(flavour, SmiFlavors.Cx3dCoordinates) && p3 != null)
{
state.atomCoords.Add(new double[] { p3.Value.X, p3.Value.Y, p3.Value.Z });
state.coordFlag = true;
}
else if (SmiFlavorTool.IsSet(flavour, SmiFlavors.CxCoordinates))
{
state.atomCoords.Add(new double[3]);
}
}
if (!state.coordFlag)
{
state.atomCoords = null;
}
// radicals
if (mol.SingleElectrons.Count > 0)
{
state.atomRads = new SortedDictionary <int, CxSmilesState.Radical>();
foreach (ISingleElectron radical in mol.SingleElectrons)
{
// 0->1, 1->2, 2->3
if (!state.atomRads.TryGetValue(EnsureNotNull(atomidx[radical.Atom]), out CxSmilesState.Radical val))
{
val = CxSmilesState.Radical.Monovalent;
}
else if (val == CxSmilesState.Radical.Monovalent)
{
val = CxSmilesState.Radical.Divalent;
}
else if (val == CxSmilesState.Radical.Divalent)
{
val = CxSmilesState.Radical.Trivalent;
}
else if (val == CxSmilesState.Radical.Trivalent)
{
throw new ArgumentException("Invalid radical state, can not be more than trivalent");
}
state.atomRads[atomidx[radical.Atom]] = val;
}
}
var sgroups = mol.GetCtabSgroups();
if (sgroups != null)
{
state.sgroups = new List <CxSmilesState.PolymerSgroup>();
state.positionVar = new SortedDictionary <int, IList <int> >();
foreach (Sgroup sgroup in sgroups)
{
switch (sgroup.Type)
{
// polymer SRU
case SgroupType.CtabStructureRepeatUnit:
case SgroupType.CtabMonomer:
case SgroupType.CtabMer:
case SgroupType.CtabCopolymer:
case SgroupType.CtabCrossLink:
case SgroupType.CtabModified:
case SgroupType.CtabMixture:
case SgroupType.CtabFormulation:
case SgroupType.CtabAnyPolymer:
case SgroupType.CtabGeneric:
case SgroupType.CtabComponent:
case SgroupType.CtabGraft:
string supscript = (string)sgroup.GetValue(SgroupKey.CtabConnectivity);
state.sgroups.Add(new CxSmilesState.PolymerSgroup(GetSgroupPolymerKey(sgroup),
ToAtomIdxs(sgroup.Atoms, atomidx),
sgroup.Subscript,
supscript));
break;
case SgroupType.ExtMulticenter:
IAtom beg = null;
List <IAtom> ends = new List <IAtom>();
ISet <IBond> bonds = sgroup.Bonds;
if (bonds.Count != 1)
{
throw new ArgumentException("Multicenter Sgroup in inconsistent state!");
}
IBond bond = bonds.First();
foreach (IAtom atom in sgroup.Atoms)
{
if (bond.Contains(atom))
{
if (beg != null)
{
throw new ArgumentException("Multicenter Sgroup in inconsistent state!");
}
beg = atom;
}
else
{
ends.Add(atom);
}
}
state.positionVar[EnsureNotNull(atomidx[beg])] =
ToAtomIdxs(ends, atomidx);
break;
case SgroupType.CtabAbbreviation:
case SgroupType.CtabMultipleGroup:
// display shortcuts are not output
break;
case SgroupType.CtabData:
// can be generated but currently ignored
break;
default:
throw new NotSupportedException("Unsupported Sgroup Polymer");
}
}
}
return(state);
}
/// <summary>
/// Creates a SMILES string of the flavour specified as a parameter
/// and write the output order to the provided array.
/// </summary>
/// <remarks>
/// The output order allows one to arrange auxiliary atom data in the
/// order that a SMILES string will be read. A simple example is seen below
/// where 2D coordinates are stored with a SMILES string. This method
/// forms the basis of CXSMILES.
/// </remarks>
/// <example>
/// <include file='IncludeExamples.xml' path='Comments/Codes[@id="NCDK.Smiles.SmilesGenerator_Example.cs+Create_IAtomContainer_int_int"]/*' />
/// </example>
/// <param name="molecule">the molecule to write</param>
/// <param name="order">array to store the output order of atoms</param>
/// <returns>the SMILES string</returns>
/// <exception cref="CDKException">a valid SMILES could not be created</exception>
public static string Create(IAtomContainer molecule, SmiFlavors flavour, int[] order)
{
try
{
if (order.Length != molecule.Atoms.Count)
{
throw new ArgumentException("the array for storing output order should be the same length as the number of atoms");
}
var g = CDKToBeam.ToBeamGraph(molecule, flavour);
// apply the canonical labelling
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.Canonical))
{
// determine the output order
var labels = Labels(flavour, molecule);
g = g.Permute(labels);
if ((flavour & SmiFlavors.AtomAtomMapRenumber) == SmiFlavors.AtomAtomMapRenumber)
{
g = Functions.RenumberAtomMaps(g);
}
if (!SmiFlavorTool.IsSet(flavour, SmiFlavors.UseAromaticSymbols))
{
g = g.Resonate();
}
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.StereoCisTrans))
{
// FIXME: required to ensure canonical double bond labelling
g.Sort(new Graph.VisitHighOrderFirst());
// canonical double-bond stereo, output be C/C=C/C or C\C=C\C
// and we need to normalise to one
g = Functions.NormaliseDirectionalLabels(g);
// visit double bonds first, prefer C1=CC=C1 to C=1C=CC1
// visit hydrogens first
g.Sort(new Graph.VisitHighOrderFirst()).Sort(new Graph.VisitHydrogenFirst());
}
var smiles = g.ToSmiles(order);
// the SMILES has been generated on a reordered molecule, transform
// the ordering
var canorder = new int[order.Length];
for (int i = 0; i < labels.Length; i++)
{
canorder[i] = order[labels[i]];
}
System.Array.Copy(canorder, 0, order, 0, order.Length);
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.CxSmilesWithCoords))
{
smiles += CxSmilesGenerator.Generate(GetCxSmilesState(flavour, molecule), flavour, null, order);
}
return(smiles);
}
else
{
string smiles = g.ToSmiles(order);
if (SmiFlavorTool.IsSet(flavour, SmiFlavors.CxSmilesWithCoords))
{
smiles += CxSmilesGenerator.Generate(GetCxSmilesState(flavour, molecule), flavour, null, order);
}
return(smiles);
}
}
catch (IOException e)
{
throw new CDKException(e.Message);
}
}
/// <summary>
/// Create a SMILES generator with the specified <see cref="SmiFlavors"/>.
/// </summary>
/// <example>
/// <include file='IncludeExamples.xml' path='Comments/Codes[@id="NCDK.Smiles.SmilesGenerator_Example.cs+ctor_SmiFlavor"]/*' />
/// </example>
/// <param name="flavour">SMILES flavour flags <see cref="SmiFlavors"/></param>
public SmilesGenerator(SmiFlavors flavour)
{
this.flavour = flavour;
}