public void Merge(ICountFingerprint fp)
{
var newFp = new Dictionary <int, int>();
{
for (int i = 0; i < hitHashes.Length; i++)
{
newFp.Add(hitHashes[i], numOfHits[i]);
}
}
{
for (int i = 0; i < fp.GetNumberOfPopulatedBins(); i++)
{
if (!newFp.TryGetValue(fp.GetHash(i), out int count))
{
count = 0;
}
newFp[fp.GetHash(i)] = count + fp.GetCount(i);
}
}
var keys = new List <int>(newFp.Keys);
keys.Sort();
hitHashes = new int[keys.Count];
numOfHits = new int[keys.Count];
{
int i = 0;
foreach (var key in keys)
{
hitHashes[i] = key;
numOfHits[i++] = newFp[key];
}
}
}
/// <summary>
/// Calculates Tanimoto distance for two count fingerprints using method 2.
/// </summary>
/// <remarks>
/// <token>cdk-cite-Grant06</token>.
/// </remarks>
/// <param name="fp1">count fingerprint 1</param>
/// <param name="fp2">count fingerprint 2</param>
/// <returns>a Tanimoto distance</returns>
public static double Method2(ICountFingerprint fp1, ICountFingerprint fp2)
{
long maxSum = 0, minSum = 0;
int i = 0, j = 0;
while (i < fp1.GetNumberOfPopulatedBins() || j < fp2.GetNumberOfPopulatedBins())
{
int?hash1 = i < fp1.GetNumberOfPopulatedBins() ? fp1.GetHash(i) : (int?)null;
int?hash2 = j < fp2.GetNumberOfPopulatedBins() ? fp2.GetHash(j) : (int?)null;
int?count1 = i < fp1.GetNumberOfPopulatedBins() ? fp1.GetCount(i) : (int?)null;
int?count2 = j < fp2.GetNumberOfPopulatedBins() ? fp2.GetCount(j) : (int?)null;
if (count2 == null || (hash1 != null && hash1 < hash2))
{
maxSum += count1.Value;
i++;
continue;
}
if (count1 == null || (hash2 != null && hash1 > hash2))
{
maxSum += count2.Value;
j++;
continue;
}
if (hash1.Equals(hash2))
{
maxSum += Math.Max(count1.Value, count2.Value);
minSum += Math.Min(count1.Value, count2.Value);
i++;
j++;
}
}
return(((double)minSum) / maxSum);
}
public void TestGetCountFingerprint()
{
var printer = new AtomPairs2DFingerprinter();
var mol1 = parser.ParseSmiles("cccccccccc");
ICountFingerprint icfp = printer.GetCountFingerprint(mol1);
Assert.AreEqual(9, icfp.GetNumberOfPopulatedBins());
}
public override void TestGetCountFingerprint()
{
var fingerprinter = new SignatureFingerprinter(0);
var sp = CDK.SmilesParser;
var mol = sp.ParseSmiles("O(NC)CC");
ICountFingerprint bitFP = fingerprinter.GetCountFingerprint(mol);
Assert.IsNotNull(bitFP);
Assert.AreNotSame(0, bitFP.Length);
}
public override void TestGetCountFingerprint()
{
var fpr = new LingoFingerprinter(4);
var sp = CDK.SmilesParser;
var mol = sp.ParseSmiles("Oc1ccccc1");
ICountFingerprint fp = fpr.GetCountFingerprint(mol);
Assert.AreEqual(2, fp.GetCountForHash("cccc".GetHashCode()));
Assert.AreEqual(1, fp.GetCountForHash("Oc0c".GetHashCode()));
Assert.AreEqual(1, fp.GetCountForHash("c0cc".GetHashCode()));
Assert.AreEqual(1, fp.GetCountForHash("0ccc".GetHashCode()));
Assert.AreEqual(1, fp.GetCountForHash("ccc0".GetHashCode()));
}
public void TestCompaRingBitFingerprintAndCountBehavingAsBit()
{
var mol1 = TestMoleculeFactory.Make123Triazole();
var mol2 = TestMoleculeFactory.MakeImidazole();
var fingerprinter = new SignatureFingerprinter(1);
ICountFingerprint countFp1 = fingerprinter.GetCountFingerprint(mol1);
ICountFingerprint countFp2 = fingerprinter.GetCountFingerprint(mol2);
countFp1.SetBehaveAsBitFingerprint(true);
countFp2.SetBehaveAsBitFingerprint(true);
var bitFp1 = fingerprinter.GetBitFingerprint(mol1);
var bitFp2 = fingerprinter.GetBitFingerprint(mol2);
var bitTanimoto = Tanimoto.Calculate(bitFp1, bitFp2);
var countTanimoto1 = Tanimoto.Method1(countFp1, countFp2);
var countTanimoto2 = Tanimoto.Method2(countFp1, countFp2);
Assert.AreEqual(countTanimoto1, countTanimoto2, 0.001);
Assert.AreEqual(bitTanimoto, countTanimoto1, 0.001);
}
/// <summary>
/// Calculates Tanimoto distance for two count fingerprints using method 1.
/// </summary>
/// <remarks>
/// The feature/count type fingerprints may be of different length.
/// Uses Tanimoto method from <token>cdk-cite-Steffen09</token>.
/// </remarks>
/// <param name="fp1">count fingerprint 1</param>
/// <param name="fp2">count fingerprint 2</param>
/// <returns>a Tanimoto distance</returns>
public static double Method1(ICountFingerprint fp1, ICountFingerprint fp2)
{
long xy = 0, x = 0, y = 0;
for (int i = 0; i < fp1.GetNumberOfPopulatedBins(); i++)
{
int hash = fp1.GetHash(i);
for (int j = 0; j < fp2.GetNumberOfPopulatedBins(); j++)
{
if (hash == fp2.GetHash(j))
{
xy += fp1.GetCount(i) * fp2.GetCount(j);
}
}
x += fp1.GetCount(i) * fp1.GetCount(i);
}
for (int j = 0; j < fp2.GetNumberOfPopulatedBins(); j++)
{
y += fp2.GetCount(j) * fp2.GetCount(j);
}
return((double)xy / (x + y - xy));
}
/// <summary>
/// BuildTest
/// </summary>
public static void BuildTest()
{
CircularFingerprinter cfp = null;
CircularFingerprinterClass FpClass = CircularFingerprinterClass.ECFP6; // FP diameter
int FpLen = 2048; // folded binary fp length
IAtomContainer mol, mol2;
//string molfile = FileUtil.ReadFile(@"C:\Download\CorpId-12345.mol");
//java.io.StringReader sr = new java.io.StringReader(molfile);
//if (Lex.Contains(molfile, "v2000"))
// cor = new MDLV2000Reader(sr);
//else
// cor = new MDLV3000Reader(sr);
//cor.setReaderMode(IChemObjectReader.Mode.RELAXED);
//ac = (IAtomContainer)cor.read(new AtomContainer());
//cor.close();
FpClass = CircularFingerprinterClass.ECFP4; // debug
cfp = new CircularFingerprinter(FpClass, FpLen);
StreamReader reader = new StreamReader(@"C:\Download\CorpId-12345.mol");
//FileReader FileReader = new FileReader(@"C:\Download\V3000 Mols.sdf");
EnumerableSDFReader rdr = new EnumerableSDFReader(reader.BaseStream, ChemObjectBuilder.Instance);
rdr.ReaderMode = ChemObjectReaderMode.Relaxed;
IEnumerator <IAtomContainer> cursor = rdr.GetEnumerator();
while (cursor.MoveNext())
{
mol = cursor.Current;
mol = CdkMol.GetLargestMoleculeFragment(mol);
ICountFingerprint cfp1 = cfp.GetCountFingerprint(mol); // get hash values and counts for each
cfp.Calculate(mol);
int fpCount = cfp.FPCount;
for (int fpi = 0; fpi < fpCount; fpi++) // gets
{
CircularFingerprint cfp2 = cfp.GetFP(fpi); // gets hash, iteration and lists of atoms (dups appear multiple times)
}
IBitFingerprint bfp = cfp.GetBitFingerprint(mol);
BitArray bs = bfp.AsBitSet();
int bsCard = bfp.Cardinality;
long bsSize = bfp.Length;
continue;
}
reader.Close();
return;
//java.io.StringReader sr = new java.io.StringReader(molfile);
//AtomContainer mol = new AtomContainer();
//mol.addAtom(new Atom("C"));
//mol.addAtom(new Atom("H"));
//mol.addAtom(new Atom("H"));
//mol.addAtom(new Atom("H"));
//mol.addAtom(new Atom("H"));
//mol.addBond(new Bond(mol.getAtom(0), mol.getAtom(1)));
//mol.addBond(new Bond(mol.getAtom(0), mol.getAtom(2)));
//mol.addBond(new Bond(mol.getAtom(0), mol.getAtom(3)));
//mol.addBond(new Bond(mol.getAtom(0), mol.getAtom(4)));
//FileReader FileReader = new FileReader(@"C:\Download\CorpId-12345.mol");
//MolReader mr = new MolReader(FileReader, DefaultChemObjectBuilder.getInstance());
//java.io.StringReader sr = new java.io.StringReader(molfile);
//IMol m = (IMol)mr.next();
//FileReader.close();
}
public void Merge(ICountFingerprint fp)
{
}
/// <summary>
/// Evaluate continuous Tanimoto coefficient for two feature, count fingerprint representations.
/// </summary>
/// <remarks>
/// Note that feature/count type fingerprints may be of different length.
/// Uses Tanimoto method from 10.1021/ci800326z
/// </remarks>
/// <param name="fp1">The first fingerprint</param>
/// <param name="fp2">The second fingerprint</param>
/// <returns>The Tanimoto coefficient</returns>
/// <seealso cref="Method1(ICountFingerprint, ICountFingerprint)"/>
/// <seealso cref="Method2(ICountFingerprint, ICountFingerprint)"/>
public static double Calculate(ICountFingerprint fp1, ICountFingerprint fp2)
{
return(Method2(fp1, fp2));
}
