/// <summary>
/// Internal - load the SMARTS patterns for each atom type from MMFFSYMB.sma.
/// </summary>
/// <param name="smaIn">input stream of MMFFSYMB.sma</param>
/// <returns>array of patterns</returns>
/// <exception cref="IOException"></exception>
internal static AtomTypePattern[] LoadPatterns(Stream smaIn)
{
var matchers = new List <AtomTypePattern>();
using (var br = new StreamReader(smaIn))
{
string line = null;
while ((line = br.ReadLine()) != null)
{
if (SkipLine(line))
{
continue;
}
var cols = Strings.Tokenize(line, ' ');
var sma = cols[0];
var symb = cols[1];
try
{
matchers.Add(new AtomTypePattern(SmartsPattern.Create(sma).SetPrepare(false), symb));
}
catch (ArgumentException ex)
{
throw new IOException(ex.Message);
}
}
return(matchers.ToArray());
}
}
/// <summary>
/// Create a pattern for the provided SMARTS - if the SMARTS is '?' a pattern
/// is not created.
/// </summary>
/// <param name="smarts">a smarts pattern</param>
/// <param name="builder">chem object builder</param>
/// <returns>the pattern to match</returns>
private static Pattern CreatePattern(string smarts, IChemObjectBuilder builder)
{
var ptrn = SmartsPattern.Create(smarts, builder);
ptrn.SetPrepare(false); // avoid redoing aromaticity etc
return(ptrn);
}
/// <summary>
/// Create a pattern for the provided SMARTS - if the SMARTS is '?' a pattern
/// is not created.
/// </summary>
/// <param name="smarts">a smarts pattern</param>
/// <returns>the pattern to match</returns>
private static Pattern CreatePattern(string smarts)
{
var ptrn = SmartsPattern.Create(smarts);
ptrn.SetPrepare(false); // avoid redoing aromaticity etc
return(ptrn);
}
/// <summary>
/// Set the SMARTS patterns.
/// </summary>
/// <param name="smarts">the SMARTS</param>
private void SetSmarts(IEnumerable <string> smarts)
{
keys.Clear();
foreach (var key in smarts)
{
var qmol = new QueryAtomContainer();
var ptrn = SmartsPattern.Create(key);
ptrn.SetPrepare(false); // prepare is done once
keys.Add(new Key(key, ptrn));
}
}
/// <summary>
/// This method calculates occurrences of the Kier & Hall E-state fragments.
/// </summary>
/// <returns>Counts of the fragments</returns>
public Result Calculate(IAtomContainer container)
{
container = AtomContainerManipulator.RemoveHydrogens(container);
var counts = new int[SMARTS.Count];
SmartsPattern.Prepare(container);
for (int i = 0; i < SMARTS.Count; i++)
{
counts[i] = SMARTS[i].MatchAll(container).CountUnique();
}
return(new Result(counts));
}
/// <summary>
/// Creat a new query pharmacophore group
/// </summary>
/// <param name="symbol">The symbol for the group</param>
/// <param name="smarts">The SMARTS pattern to be used for matching</param>
public PharmacophoreQueryAtom(string symbol, string smarts)
{
this.Symbol = symbol;
this.Smarts = smarts;
// Note that we allow a special form of SMARTS where the | operator
// represents logical or of multi-atom groups (as opposed to ','
// which is for single atom matches)
var subSmarts = Strings.Tokenize(smarts, '|');
CompiledSmarts = new SmartsPattern[subSmarts.Count];
for (int i = 0; i < CompiledSmarts.Length; i++)
{
CompiledSmarts[i] = SmartsPattern.Create(subSmarts[i]).SetPrepare(false);
}
}
/// <inheritdoc/>
public override IBitFingerprint GetBitFingerprint(IAtomContainer atomContainer)
{
int bitsetLength = PATTERNS.Count;
BitArray fingerPrint = new BitArray(bitsetLength);
SmartsPattern.Prepare(atomContainer);
for (int i = 0; i < PATTERNS.Count; i++)
{
if (PATTERNS[i].Matches(atomContainer))
{
fingerPrint.Set(i, true);
}
}
return(new BitSetFingerprint(fingerPrint));
}
/// <inheritdoc/>
public override ICountFingerprint GetCountFingerprint(IAtomContainer atomContainer)
{
if (!keys.Any())
{
throw new CDKException("No substructures were defined");
}
// init SMARTS invariants (connectivity, degree, etc)
SmartsPattern.Prepare(atomContainer);
var map = new SortedDictionary <int, int>();
for (int i = 0; i < keys.Count; i++)
{
var ptrn = keys[i].Pattern;
map[i] = ptrn.MatchAll(atomContainer).CountUnique();
}
return(new CountFingerprint(map));
}
/// <inheritdoc/>
public override IBitFingerprint GetBitFingerprint(IAtomContainer atomContainer)
{
if (!keys.Any())
{
throw new CDKException("No substructures were defined");
}
SmartsPattern.Prepare(atomContainer);
var fingerPrint = new BitArray(keys.Count);
for (int i = 0; i < keys.Count; i++)
{
if (keys[i].Pattern.Matches(atomContainer))
{
fingerPrint[i] = true;
}
}
return(new BitSetFingerprint(fingerPrint));
}
private void PrepareInput(IAtomContainer input)
{
SmartsPattern.Prepare(input);
}
/// <inheritdoc/>
public override IBitFingerprint GetBitFingerprint(IAtomContainer container)
{
var keys = GetKeys(container.Builder);
var fp = new BitArray(keys.Count);
// init SMARTS invariants (connectivity, degree, etc)
SmartsPattern.Prepare(container);
int numAtoms = container.Atoms.Count;
var bmap = EdgeToBondMap.WithSpaceFor(container);
var adjlist = GraphUtil.ToAdjList(container, bmap);
for (int i = 0; i < keys.Count; i++)
{
var key = keys[i];
var pattern = key.Pattern;
switch (key.Smarts)
{
case "[!*]":
break;
case "[!0]":
foreach (IAtom atom in container.Atoms)
{
if (atom.MassNumber != null)
{
fp.Set(i, true);
break;
}
}
break;
// ring bits
case "[R]1@*@*@1": // 3M RING bit22
case "[R]1@*@*@*@1": // 4M RING bit11
case "[R]1@*@*@*@*@1": // 5M RING bit96
case "[R]1@*@*@*@*@*@1": // 6M RING bit163, x2=bit145
case "[R]1@*@*@*@*@*@*@1": // 7M RING, bit19
case "[R]1@*@*@*@*@*@*@*@1": // 8M RING, bit101
// handled separately
break;
case "(*).(*)":
// bit 166 (*).(*) we can match this in SMARTS but it's faster to just
// count the number of components or in this case try to traverse the
// component, iff there are some atoms not visited we have more than
// one component
bool[] visit = new bool[numAtoms];
if (numAtoms > 1 && VisitPart(visit, adjlist, 0, -1) < numAtoms)
{
fp.Set(165, true);
}
break;
default:
if (key.Count == 0)
{
if (pattern.Matches(container))
{
fp.Set(i, true);
}
}
else
{
// check if there are at least 'count' unique hits, key.count = 0
// means find at least one match hence we add 1 to out limit
if (pattern.MatchAll(container).GetUniqueAtoms().AtLeast(key.Count + 1))
{
fp.Set(i, true);
}
}
break;
}
}
// Ring Bits
// threshold=126, see AllRingsFinder.Threshold.PubChem_97
if (numAtoms > 2)
{
AllCycles allcycles = new AllCycles(adjlist,
Math.Min(8, numAtoms),
126);
int numArom = 0;
foreach (int[] path in allcycles.GetPaths())
{
// length is +1 as we repeat the closure vertex
switch (path.Length)
{
case 4: // 3M bit22
fp.Set(21, true);
break;
case 5: // 4M bit11
fp.Set(10, true);
break;
case 6: // 5M bit96
fp.Set(95, true);
break;
case 7: // 6M bit163->bit145, bit124 numArom > 1
if (numArom < 2)
{
if (IsAromPath(path, bmap))
{
numArom++;
if (numArom == 2)
{
fp.Set(124, true);
}
}
}
if (fp[162])
{
fp.Set(144, true); // >0
}
else
{
fp.Set(162, true); // >1
}
break;
case 8: // 7M bit19
fp.Set(18, true);
break;
case 9: // 8M bit101
fp.Set(100, true);
break;
}
}
}
return(new BitSetFingerprint(fp));
}