Converts a list of SMILES strings to a BatchMolGraph containing the batch of molecular graphs. :param smiles_batch: A list of SMILES strings

Returns the RDKit molecules associated with the molecules. :return: A list of RDKit Mols.

Trains a model and returns test scores on the model checkpoint with the highest validation score. :param args: Arguments. :param logger:

Saves a model checkpoint. :param model: A MoleculeModel. :param scaler: A StandardScaler fitted on the data. :param features_scaler:

Sets the features of the molecule. :param features: A 1-D numpy array of features for the molecule.

Saving model w/ its params. Get rid of the ONNX Protocal. F-string feature new in Python 3.6+ is used.

Transforms the data by subtracting the means and dividing by the standard deviations. :param X: A list of lists of floats. :

Walks the checkpoint directory to find all checkpoints, updating args.checkpoint_paths and args.ensemble_size. :param args: Arguments.

Create a Vocabulary given Counter. Args: counter: collections.Counter object holding the frequencies of e

Take labels of neighbors, sort them lexicographically and join.

Perform Weisfeiler-Lehman relabeling of the vertices

Adds predict arguments to an ArgumentParser. :param parser: An ArgumentParser.

Adds training arguments to an ArgumentParser. :param parser: An ArgumentParser.

Compute 'Scaled Dot Product Attention

Builds a learning rate scheduler. :param optimizer: The Optimizer whose learning rate will be scheduled. :param args: Arguments. :pa

Builds an Optimizer. :param model: The model to optimize. :param args: Arguments. :return: An initialized Optimizer.

Split a dataset by cluster so that no molecules in a same cluster are in the same split. :param data: A MoleculeDataset. :param sizes: A

Computes the metric based on it's accumulated state. This is called at the end of each epoch. Returns: Any: the a

Computes the norm of the gradients of a model.

Computes the norm of the parameters of a model.

Create a logger. Use a different log file for each process.

Evaluates an ensemble of models on a dataset. :param model: A model. :param data: A MoleculeDataset. :param num_tasks: Number of tas

Extract random walks of depth - 1 hops rooted in root.

Returns the size of the features array associated with each molecule. :return: The size of the features.

Compute the Bemis-Murcko scaffold for a SMILES string. :param mol: A smiles string or an RDKit molecule. :param include_chirality: Wheth

Computes (if necessary) and returns a mapping from each atom index to all neighboring atom indices. :return: A PyTorch tensor contai

Determines the proportions of the different classes in the classification dataset. :param data: A classification dataset :return: A list

Create a directory to store the experiment.

Get all the neighbors of vertex (vertex -> neighbor).

Gets the loss function corresponding to a given dataset type. :param args: Namespace containing the dataset type ("classification" or "regre

Gets a list of the current learning rates.

orthogonal init yields generally good results than uniform init

Loads features saved in a variety of formats. Supported formats: - .npz compressed (assumes features are saved with name "features")

Loads the scalers a model was trained with. :param path: Path where model checkpoint is saved. :return: A tuple with the data scaler and

Logs and returns statistics about counts and average target values in molecular clusters. :param data: A MoleculeDataset. :param index_s

Makes predictions. If smiles is provided, makes predictions on smiles. Otherwise makes predictions on args.test_data. :param args: Arguments

Modifies and validates predicting args in place. :param args: Arguments.

German and Frence have different vowels than English. This utilization removes all the non-unicode characters. Example: āáǎà -->

Trains a model and returns test scores on the model checkpoint with the highest validation score. :param args: Arguments. :param logger:

Convert a rdflib.Graph (located at file) to our KnowledgeGraph.

k-time independent runs

Split a dataset by scaffold so that no molecules sharing a scaffold are in the same split. :param data: A MoleculeDataset. :param sizes:

Computes scaffold for each smiles string and returns a mapping from scaffolds to sets of smiles. :param mols: A list of smiles strings or RD

Sets the targets for each molecule in the dataset. Assumes the targets are aligned with the datapoints. :param targets: A list of li

Trains a model for an epoch. :param model: Model. :param data: A MoleculeDataset (or a list of MoleculeDatasets if using moe). :para

Construct a feature vector for the provided instances. Parameters ---------- graphs: graph.KnowledgeGraph The gra

Looks up indices of text tokens according to the vocabulary. Parameters ---------- tokens : str or list of strs A

Gets one or more MoleculeDatapoints via an index or slice. :param item: An index (int) or a slice object. :return: A Molecul