chemfp.cdk_types module

This module should not be imported directly.

It contains internal implementation details of CDK fingerprint generation.

This module is included in the documentation because parts of this module are returned to the user, and are part of the public API.

class chemfp.cdk_types.CDKAtomPairs2DFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the atom pairs fingerprint from Yap Chun Wei’s PaDEL.

The CDK-AtomPairs2D/2.0 FingerprintType takes no parameters.

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-AtomPairs2D/2.0'

the fingerprint name

num_bits: int = 780
class chemfp.cdk_types.CDKBaseFingerprintType(fingerprint_kwargs)

Bases: ThreadsafeFingerprinterMixin, FingerprintType

fingerprinter_can_fail: bool = True

a CDK fingerprinter can raise an exception if the molecule isn’t correctly prepared

from_inchi(content: str | bytes, *, delimiter: Literal['to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', ' ', '\t'] | None = 'to-eol', errors: str = 'strict')

Generate a fingerprint from an InChI string and its id

This is equivalent to calling:

mol = fptype.toolkit.parse_inchi(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • delimiter (One of None, 'to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', or the space or tab characters (default: "to-eol")) – The separator between the SMILES and the id

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_inchistring(content: str | bytes, *, delimiter: Literal['to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', ' ', '\t'] | None = 'to-eol', errors: str = 'strict')

Generate a fingerprint from an InChI string

This is equivalent to calling:

mol = fptype.toolkit.parse_inchistring(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • delimiter (One of None, 'to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', or the space or tab characters (default: "to-eol")) – The separator between the SMILES and the id

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_molfile(content: str | bytes, *, AddStereo0d: bool = True, AddStereoElements: bool = True, InterpretHydrogenIsotopes: bool = True, ForceReadAs3DCoordinates: bool = False, mode: Literal['RELAXED', 'STRICT'] = 'RELAXED', hydrogens: Literal['as-is', 'make-explicit', 'make-implicit', 'make-nonchiral-implicit'] = 'as-is', implementation: Literal['cdk', 'chemfp'] | None = 'cdk', errors: str = 'strict')

Generate a fingerprint from a molfile record

This is equivalent to calling:

mol = fptype.toolkit.parse_molfile(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • AddStereo0d (Boolean (default: True)) – if true, create stereo from parity value when no coordinates

  • AddStereoElements (Boolean (default: True)) – if true, detect and create IStereoElements

  • InterpretHydrogenIsotopes (Boolean (default: True)) – if true, interpret D and T as hydrogen isotopes

  • ForceReadAs3DCoordinates (Boolean (default: False)) – if true, always interpret coordinates as 3D

  • mode ('RELAXED' will attempt to recover, 'STRICT' will not) – strictness mode when parsing a record

  • hydrogens (One of 'as-is', 'make-explicit', 'make-implicit', or 'make-nonchiral-implicit' (default: "as-is")) – Specify how to handle implicit or explicit hydrogens

  • implementation (either 'cdk' or 'chemfp') – use CDK or chemfp to identify records

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_sdf(content: str | bytes, *, AddStereo0d: bool = True, AddStereoElements: bool = True, InterpretHydrogenIsotopes: bool = True, ForceReadAs3DCoordinates: bool = False, mode: Literal['RELAXED', 'STRICT'] = 'RELAXED', hydrogens: Literal['as-is', 'make-explicit', 'make-implicit', 'make-nonchiral-implicit'] = 'as-is', implementation: Literal['cdk', 'chemfp'] | None = 'cdk', errors: str = 'strict')

Generate a fingerprint from an SDF record

This is equivalent to calling:

mol = fptype.toolkit.parse_sdf(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • AddStereo0d (Boolean (default: True)) – if true, create stereo from parity value when no coordinates

  • AddStereoElements (Boolean (default: True)) – if true, detect and create IStereoElements

  • InterpretHydrogenIsotopes (Boolean (default: True)) – if true, interpret D and T as hydrogen isotopes

  • ForceReadAs3DCoordinates (Boolean (default: False)) – if true, always interpret coordinates as 3D

  • mode ('RELAXED' will attempt to recover, 'STRICT' will not) – strictness mode when parsing a record

  • hydrogens (One of 'as-is', 'make-explicit', 'make-implicit', or 'make-nonchiral-implicit' (default: "as-is")) – Specify how to handle implicit or explicit hydrogens

  • implementation (either 'cdk' or 'chemfp') – use CDK or chemfp to identify records

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_smi(content: str | bytes, *, has_header: bool = False, delimiter: Literal['to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', ' ', '\t'] | None = 'to-eol', implementation: Literal['cdk', 'chemfp'] | None = 'cdk', kekulise: bool = True, cxsmiles: bool = True, hydrogens: Literal['as-is', 'make-explicit', 'make-implicit', 'make-nonchiral-implicit'] = 'as-is', errors: str = 'strict')

Generate a fingerprint from a SMILES string and its id

This is equivalent to calling:

mol = fptype.toolkit.parse_smi(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • has_header (Boolean (default: False)) – If true, treat the first line of the SMILES file as a header

  • delimiter (One of None, 'to-eol', 'space', 'tab', 'comma', 'whitespace', 'native', or the space or tab characters (default: "to-eol")) – The separator between the SMILES and the id

  • implementation (either 'cdk' or 'chemfp') – use CDK or chemfp to identify records

  • kekulise (Boolean (default: True)) – if true, ensure a valid Kekule intepretation exists

  • cxsmiles (Boolean (default: True)) – If true, look for ChemAxon CXSMILES extensions after the SMILES string (only works with ‘chemfp’ implementation)

  • hydrogens (One of 'as-is', 'make-explicit', 'make-implicit', or 'make-nonchiral-implicit' (default: "as-is")) – Specify how to handle implicit or explicit hydrogens

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_smiles(content: str | bytes, *, kekulise: bool = True, cxsmiles: bool = True, hydrogens: Literal['as-is', 'make-explicit', 'make-implicit', 'make-nonchiral-implicit'] = 'as-is', errors: str = 'strict')

Generate a fingerprint from a SMILES string

This is equivalent to calling:

mol = fptype.toolkit.parse_smistring(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • kekulise (Boolean (default: True)) – if true, ensure a valid Kekule intepretation exists

  • cxsmiles (Boolean (default: True)) – If true, look for ChemAxon CXSMILES extensions after the SMILES string

  • hydrogens (One of 'as-is', 'make-explicit', 'make-implicit', or 'make-nonchiral-implicit' (default: "as-is")) – Specify how to handle implicit or explicit hydrogens

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

from_smistring(content: str | bytes, *, kekulise: bool = True, cxsmiles: bool = True, hydrogens: Literal['as-is', 'make-explicit', 'make-implicit', 'make-nonchiral-implicit'] = 'as-is', errors: str = 'strict')

Generate a fingerprint from a SMILES string

This is equivalent to calling:

mol = fptype.toolkit.parse_smistring(content, ..., errors=errors)
fp = fptype.from_mol(mol) if (mol is not None) else None
Parameters:

  • kekulise (Boolean (default: True)) – if true, ensure a valid Kekule intepretation exists

  • cxsmiles (Boolean (default: True)) – If true, look for ChemAxon CXSMILES extensions after the SMILES string

  • hydrogens (One of 'as-is', 'make-explicit', 'make-implicit', or 'make-nonchiral-implicit' (default: "as-is")) – Specify how to handle implicit or explicit hydrogens

  • errors (one of "strict", "ignore", or "log") – specify how to handle errors

Returns:

a fingerprint byte string

module = <module 'chemfp.cdk_toolkit>'
software: _OptionalStr = ...

a description of the CDK and chemfp versions

class chemfp.cdk_types.CDKDaylightFingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s Daylight-like fingerprint, version 2.0

The CDK-Daylight/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • searchDepth - maximum path length (default: 7)

  • pathLimit - maximum number of paths (default: 42000)

  • hashPseudoAtoms - include pseudo-atoms in the hash calculation (default: 0)

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-Daylight/2.0'

the fingerprint name

class chemfp.cdk_types.CDKECFP0FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the ECFP0 fingerprint, version 2.0

The CDK-ECFP0/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-ECFP0/2.0'

the fingerprint name

class chemfp.cdk_types.CDKECFP2FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the ECFP2 fingerprint, version 2.0

The CDK-ECFP2/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-ECFP2/2.0'

the fingerprint name

class chemfp.cdk_types.CDKECFP4FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the ECFP4 fingerprint, version 2.0

The CDK-ECFP4/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-ECFP4/2.0'

the fingerprint name

class chemfp.cdk_types.CDKECFP6FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the ECFP6 fingerprint, version 2.0

The CDK-ECFP6/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-ECFP6/2.0'

the fingerprint name

class chemfp.cdk_types.CDKEStateFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the EState fingerprint, version 2.0

The CDK-EState/2.0 fingerprints have no parameters.

This fingerprinter requires fully implicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-implicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-EState/2.0'

the fingerprint name

num_bits: int = 79
class chemfp.cdk_types.CDKExtendedFingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s Extended fingerprint, version 2.0

This extends the DaylightFingerprinter with 25 additional bits describing ring features and isotopic masses.

The CDK-Extended/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • searchDepth - maximum path length (default: 7)

  • pathLimit - maximum number of paths (default: 42000)

  • hashPseudoAtoms - include pseudo-atoms in the hash calculation (default: 0)

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-Extended/2.0'

the fingerprint name

class chemfp.cdk_types.CDKFCFP0FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the FCFP0 fingerprint, version 2.0

The CDK-FCFP0/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-FCFP0/2.0'

the fingerprint name

class chemfp.cdk_types.CDKFCFP2FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the FCFP2 fingerprint, version 2.0

The CDK-FCFP2/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-FCFP2/2.0'

the fingerprint name

class chemfp.cdk_types.CDKFCFP4FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the FCFP4 fingerprint, version 2.0

The CDK-FCFP4/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-FCFP4/2.0'

the fingerprint name

class chemfp.cdk_types.CDKFCFP6FingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s implementation of the FCFP6 fingerprint, version 2.0

The CDK-FCFP6/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • perceiveStereochemistry - if True, re-perceive stereochemistry from 2D/3D coordinates

name: str = 'CDK-FCFP6/2.0'

the fingerprint name

class chemfp.cdk_types.CDKGraphOnlyFingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s GraphOnly fingerprint, version 2.0

This is a variant of the DaylightFingerprinter which does not take bond orders into account.

The CDK-GraphOnly/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • searchDepth - maximum path length (default: 7)

  • pathLimit - maximum number of paths (default: 42000)

  • hashPseudoAtoms - include pseudo-atoms in the hash calculation (default: 0)

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-GraphOnly/2.0'

the fingerprint name

class chemfp.cdk_types.CDKHybridizationFingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s Hybridization fingerprint, version 2.0

This is a variant of the DaylightFingerprinter which ignores aromaticity but instead takes into account SP2 hybridization.

The CDK-Hybridization/2.0 FingerprintType parameters are:

  • size - the number of bits in the fingerprint (default: 1024)

  • searchDepth - maximum path length (default: 7)

  • pathLimit - maximum number of paths (default: 42000)

  • hashPseudoAtoms - include pseudo-atoms in the hash calculation (default: 0)

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-Hybridization/2.0'

the fingerprint name

class chemfp.cdk_types.CDKKlekotaRothFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the Klekota-Roth fingerprints

The CDK-KlekotaRoth/2.0 fingerprints have no parameters.

name: str = 'CDK-KlekotaRoth/2.0'

the fingerprint name

num_bits: int = 4860
class chemfp.cdk_types.CDKMACCSFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the 166 MACCS keys, version 2.0

The CDK-MACCS/2.0 fingerprints have no parameters.

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-MACCS/2.0'

the fingerprint name

num_bits: int = 166
class chemfp.cdk_types.CDKPubchemFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the PubChem fingerprints

This is generated by CDK versions up to and including 2.8.

The CDK-Pubchem/2.0 fingerprints have no parameters.

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

Note that the PubChem fingerprinter in CDK 2.9 and later does not require explicit hydrogens.

name: str = 'CDK-Pubchem/2.0'

the fingerprint name

num_bits: int = 881
class chemfp.cdk_types.CDKPubchemFingerprintType_v29(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s implementation of the PubChem fingerprints

This is generated by CDK versions 2.9 and later. Quoting the release notes:

“Explicit hydrogens are not longer required and there is an option to use a more correct ring set definition matching closer the original CACTVS substructure keys.”

The CDK-Pubchem/2.9 fingerprints have no parameters.

(The underlying CDK fingerprinter has an option, ‘esssr’, to use the older and less correct ring set definition, but I cannot think of a reason for chemfp to support it, even after dicussion with one of the CDK authors.)

name: str = 'CDK-Pubchem/2.9'

the fingerprint name

num_bits: int = 881
class chemfp.cdk_types.CDKShortestPathFingerprintType_v20(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s ShortestPath fingerprint, version 2.0

Calculate a fingerprint based on the shortest paths between two atoms.

The CDK-ShortestPath/2.0 FingerprintType parameter is:

  • size - the number of bits in the fingerprint (default: 1024)

This is generated by CDK versions older than 2.7.

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

fingerprinter_can_fail: bool = True

a CDK fingerprinter can raise an exception if the molecule isn’t correctly prepared

name: str = 'CDK-ShortestPath/2.0'

the fingerprint name

class chemfp.cdk_types.CDKShortestPathFingerprintType_v27(fingerprint_kwargs)

Bases: VariableSizeFingerprint

CDK’s ShortestPath fingerprint, version 2.7

Calculate a fingerprint based on the shortest paths between two atoms.

The CDK-ShortestPath/2.7 FingerprintType parameter is:

  • size - the number of bits in the fingerprint (default: 1024)

This version is new in CDK 2.7, where the internal PRNG was changed from a Mersenne Twister to XorShift, resulting in a different bit pattern.

This fingerprinter requires explicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-explicit” to convert any implicit hydrogens to explicit.

fingerprinter_can_fail: bool = True

a CDK fingerprinter can raise an exception if the molecule isn’t correctly prepared

name: str = 'CDK-ShortestPath/2.7'

the fingerprint name

class chemfp.cdk_types.CDKSubstructureFingerprintType_v20(fingerprint_kwargs: Dict[str, Any])

Bases: NoFingerprintParametersMixin, FixedSizeFingerprint

CDK’s Substructure key fingerprints

The CDK-Substructure/2.0 fingerprints have no parameters.

This fingerprinter requires fully implicit hydrogens. When using chemfp to read a SMILES or SDF, set the “hydrogens” reader argument to “make-implicit” to convert any implicit hydrogens to explicit.

name: str = 'CDK-Substructure/2.0'

the fingerprint name

num_bits: int = 307