e3fp.fingerprint.metrics package

Submodules

Module contents

Efficient comparison metrics for fingerprints and their databases.

Author: Seth Axen E-mail: seth.axen@gmail.com

cosine(A, B=None)[source]

Compute cosine similarities between fingerprints.

Fingerprints must have same number of bits. If only one fingerprint/database is provided, it is compared to self.

Parameters

A, B (Fingerprint or FingerprintDatabase) – Fingerprint(s) to be compared

Returns

cosine

Return type

float or ndarray [shape (num_fps_A, num_fps_B)]

dice(A, B=None)[source]

Compute Dice coefficients between fingerprints.

Fingerprints must have same number of bits. If not bit-fingerprints, arrays will be cast to binary. If only one fingerprint/database is provided, it is compared to self.

Parameters

A, B (Fingerprint or FingerprintDatabase) – Fingerprint(s) to be compared

Returns

dice

Return type

float or ndarray [shape (num_fps_A, num_fps_B)]

pearson(A, B=None)[source]

Compute Pearson correlation between fingerprints.

Fingerprints must have same number of bits. If only one fingerprint/database is provided, it is compared to self.

Parameters

A, B (Fingerprint or FingerprintDatabase) – Fingerprint(s) to be compared

Returns

pearson

Return type

float or ndarray [shape (num_fps_A, num_fps_B)]

See also

cosine, dice, soergel, tanimoto

soergel(A, B=None)[source]

Compute Soergel similarities between fingerprints.

Soergel similarity is the complement of the Soergel distance and is analogous to the Tanimoto coefficient for count/float fingerprints. For binary data, it is equivalent to tanimoto.

Parameters

A, B (Fingerprint or FingerprintDatabase) – Fingerprint(s) to be compared

Returns

soergel

Return type

float or ndarray [shape (num_fps_A, num_fps_B)]

See also

cosine, dice, pearson, tanimoto

tanimoto(A, B=None)[source]

Compute Tanimoto coefficients between fingerprints.

Fingerprints must have same number of bits. If not bit-fingerprints, arrays will be cast to binary. For non-binary data, use soergel. If only one fingerprint/database is provided, it is compared to self.

Parameters

A, B (Fingerprint or FingerprintDatabase) – Fingerprint(s) to be compared

Returns

tanimoto – Pairwise tanimoto(s) between fingerprint(s) in A and B.

Return type

float or ndarray [shape (num_fps_A, num_fps_B)]

See also

cosine, dice, pearson, soergel