CLAMPFingerprint#

class skfp.fingerprints.CLAMPFingerprint(weights_path: str | None = None, n_jobs: int | None = None, batch_size: int | None = None, verbose: int | dict = 0)#

CLAMP fingerprint.

CLAMP (Contrastive Language And Molecule Pre-training) uses a pretrained two-layer MLP compound encoder from CLAMP [1] to transform concatenated ECFP count (4096 bits) and RDKit count (4096 bits) fingerprints into 768-dimensional learned embeddings.

Requires neural optional dependency, installed as scikit-fingerprints[neural]

Parameters:

  • weights_path (str or None, default=None) – Path to a local pretrained checkpoint file (.pt). If None, weights are downloaded automatically from the scikit-fingerprints/clamp HuggingFace Hub repository and cached in the standard HuggingFace cache directory (~/.cache/huggingface/hub/ by default).

  • n_jobs (int, default=None) – The number of jobs to run in parallel. transform() is parallelized over the input molecules. None means 1 unless in a joblib.parallel_backend context. -1 means using all processors. See scikit-learn documentation on n_jobs for more details.

  • batch_size (int, default=None) – Number of inputs processed in each batch. None divides input data into equal-sized parts, as many as n_jobs.

  • verbose (int or dict, default=0) – Controls the verbosity when computing fingerprints. If a dictionary is passed, it is treated as kwargs for tqdm(), and can be used to control the progress bar.

n_features_out#

Number of output features, i.e. the CLAMP embedding dimension.

Type:

int = 768

requires_conformers#

This fingerprint uses only 2D molecular graphs and does not require conformers.

Type:

bool = False

References

Examples

>>> from skfp.fingerprints.neural import CLAMPFingerprint
>>> smiles = ["O", "CC", "[C-]#N", "CC=O"]
>>> fp = CLAMPFingerprint()
>>> fp.transform(smiles)  
array([...], shape=(4, 768), dtype=float32)

Methods

fit(X[, y])

Unused, kept for scikit-learn compatibility.

fit_transform(X[, y])

The same as .transform() method, kept for scikit-learn compatibility.

get_feature_names_out([input_features])

Get output feature names for transformation.

get_input_features(X)

Compute CLAMP encoder input features.

get_metadata_routing()

Get metadata routing of this object.

get_model()

Return the pretrained CLAMP compound encoder.

get_params([deep])

Get parameters for this estimator.

set_output(*[, transform])

Set output container.

set_params(**params)

Set the parameters of this estimator.

set_transform_request(*[, copy])

Configure whether metadata should be requested to be passed to the transform method.

transform(X[, copy])

Compute CLAMP fingerprints.
fit(X: Sequence[str | Mol], y: Any | None = None, **fit_params)#

Unused, kept for scikit-learn compatibility.

Parameters:

  • X (any) – Unused, kept for scikit-learn compatibility.

  • y (any) – Unused, kept for scikit-learn compatibility.

  • **fit_params (dict) – Unused, kept for scikit-learn compatibility.

Return type:

self

fit_transform(X: Sequence[str | Mol], y: Any | None = None, **fit_params)#

The same as .transform() method, kept for scikit-learn compatibility.

Parameters:

  • X (any) – See .transform() method.

  • y (any) – See .transform() method.

  • **fit_params (dict) – Unused, kept for scikit-learn compatibility.

Returns:

X_new – See .transform() method.

Return type:

any

get_feature_names_out(input_features=None)#

Get output feature names for transformation.

The feature names out will prefixed by the lowercased class name. For example, if the transformer outputs 3 features, then the feature names out are: [“class_name0”, “class_name1”, “class_name2”].

Parameters:

input_features (array-like of str or None, default=None) – Only used to validate feature names with the names seen in fit.

Returns:

feature_names_out – Transformed feature names.

Return type:

ndarray of str objects

get_input_features(X: Sequence[str | Mol]) → ndarray#

Compute CLAMP encoder input features.

Returns the intermediate 8192-dimensional representation used as input to the pretrained encoder in transform(): a log-scaled sum of ECFP count and RDKit count fingerprints.

Parameters:

X ({sequence of str or Mol}) – Sequence containing SMILES strings or RDKit Mol objects.

Returns:

X – Array with encoder input features as float32.

Return type:

ndarray of shape (n_samples, 8192)

get_metadata_routing()#

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:

routing – A MetadataRequest encapsulating routing information.

Return type:

MetadataRequest

get_model() → CLAMPCompoundEncoder#

Return the pretrained CLAMP compound encoder.

Returns:

encoder – Pretrained nn.Module in eval mode.

Return type:

CLAMPCompoundEncoder

get_params(deep=True)#

Get parameters for this estimator.

Parameters:

deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:

params – Parameter names mapped to their values.

Return type:

dict

set_output(*, transform=None)#

Set output container.

See Introducing the set_output API for an example on how to use the API.

Parameters:

transform ({"default", "pandas", "polars"}, default=None) –

Configure output of transform and fit_transform.

  • ”default”: Default output format of a transformer

  • ”pandas”: DataFrame output

  • ”polars”: Polars output

  • None: Transform configuration is unchanged

Added in version 1.4: “polars” option was added.

Returns:

self – Estimator instance.

Return type:

estimator instance

set_params(**params)#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:

**params (dict) – Estimator parameters.

Returns:

self – Estimator instance.

Return type:

estimator instance

set_transform_request(*, copy: bool | None | str = '$UNCHANGED$') → CLAMPFingerprint#

Configure whether metadata should be requested to be passed to the transform method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to transform if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to transform.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

copy (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for copy parameter in transform.

Returns:

self – The updated object.

Return type:

object

transform(X: Sequence[str | Mol], copy: bool = False) → ndarray#

Compute CLAMP fingerprints.

Parameters:

  • X ({sequence of str or Mol}) – Sequence containing SMILES strings or RDKit Mol objects.

  • copy (bool, default=False) – Whether to copy input data.

Returns:

X – Array with CLAMP embeddings as float32.

Return type:

ndarray of shape (n_samples, 768)