chemaxon.descriptors
Class CFGenerator

java.lang.Object
  chemaxon.descriptors.MDGenerator
      chemaxon.descriptors.CFGenerator

public class CFGenerator
extends chemaxon.descriptors.MDGenerator

The CFGenerator class generates topological fingerprints of molecular graphs.

Basic concepts
A binary string (series of 0 and 1) is constructed based on local connectivity atoms and atom types. The length of the series is a predefined constant parameter. Two further parameters influence the fingerprints gerenerated: number of bonds, which determines the number size of local neighborhood of atoms takes into account by providing an upper limit for the length of path starting from each atom; and the number of bits to be set in the fingerprint for each property identified.

Typical usage
For the sake of optimal memory usage one instance of this class can generate fingerprints for a series of molecular graphs by the consequtive call to the generate() method.
In most cases the generator is not intended to be used directly. When molecules are taken from files or databases the corresponding MolecularDescriptors can be generated by the appropriate MDReader object. Alternatively, MolecularDescriptor.generate( final Molecule ) is the simplest way to obtain a descriptor corresponding to a molecular structure.

Example of the direct use of the class within an application:

      CFGenerator gen = new ChemicalFingerprintenerator( 
              new Standardizer( new FileInputStream("stdconfig.xml") ), 512 );
      ChemicalFingerprint fp = new ChemicalFingerprint();
      Molecule mol = getFirstMoleculeFromSomewhere();
      while ( mol != null ) {
          gen.generate( mol, fp );
          doSomethingWith( fp );
          mol = getNextMoleculeFromSomewhere();
      }
 

Since:

JChem 2.0

Author:

Miklos Vargyas

Field Summary

Fields inherited from class chemaxon.descriptors.MDGenerator

createStatistics, density, freqCount, maxNonEmptyId, maxNonEmptyPercent, minNonEmptyId, minNonEmptyPercent, molCount, sumNonEmptyPercent

Constructor Summary

CFGenerator(int length)
Creates a new instance of CFGenerator which can be used to generate chemical fingerprints for an arbitrary number of molecules.

CFGenerator(Standardizer s, int length)
Deprecated. since 2.2

Method Summary

protected int	calcFreqCount(MolecularDescriptor d) Updates statistics gathered on fingerprints generated and get the number of non-zero cells.
java.lang.String[]	generate(Molecule m, MolecularDescriptor d) Generates the chemical fingerprint for the given molecule.

Methods inherited from class chemaxon.descriptors.MDGenerator

getAverageNonZeroRatio, getBrightestMolId, getDarkestMolId, getDensityCounts, getFrequencyCounts, getMaximumBitRatio, getMinimumBitRatio, getMoleculeCount, setCreateStatistics, updateStatistics

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CFGenerator

public CFGenerator(Standardizer s,
                   int length)

Deprecated. since 2.2

Creates a new instance of CFGenerator which can be used to generate chemical fingerprints for an arbitrary number of molecules.

Parameters:

s - Standardizer object to transform input molecules into a standard form

length - length of the chemical fingerprint in bits

CFGenerator

public CFGenerator(int length)

Creates a new instance of CFGenerator which can be used to generate chemical fingerprints for an arbitrary number of molecules.

Parameters:

length - length of the chemical fingerprint in bits

Since:

JChem 2.2

Method Detail

generate

public java.lang.String[] generate(Molecule m,
                                   MolecularDescriptor d)
                            throws chemaxon.descriptors.MDGeneratorException

Generates the chemical fingerprint for the given molecule. New instance of the ChemicalFingerprint object is not allocated, the MolecularDescriptor provided as a method parameter is updated (and it has to be allocated and initialized by the client of this class).

Specified by:

generate in class chemaxon.descriptors.MDGenerator

Parameters:

m - molecule for which the fingerprint is created

d - the chemical fingerprint generated

Returns:

names of tags (properties) set (added) during fingerprint generation, it is always null in the case of ChemicalFingerprint

Throws:

chemaxon.descriptors.MDGeneratorException - in the case of any failures to generate the descriptor

calcFreqCount

protected int calcFreqCount(MolecularDescriptor d)

Updates statistics gathered on fingerprints generated and get the number of non-zero cells.

Overrides:

calcFreqCount in class chemaxon.descriptors.MDGenerator

Parameters:

d - newly generated MolecularDescriptor

Returns:

brightness of the fingerprint

Since:

JChem 2.1