qsprpred.extra.gpu.models package

property classPath: str

Return the fully classified path of the model.

Returns:: class path of the model
Return type:: str

cleanFiles()

Clean up the model files.

Removes the model directory and all its contents.

Convert the given data matrix and target matrix to np.ndarray format.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix

Returns:

data matrix and/or target matrix in np.ndarray format

createPredictionDatasetFromMols(mols: list[str | rdkit.Chem.rdchem.Mol], smiles_standardizer: str | Callable[[str], str] = 'chembl', n_jobs: int = 1, fill_value: float = nan) → tuple[qsprpred.data.tables.qspr.QSPRDataset, numpy.ndarray]

Create a QSPRDataset instance from a list of SMILES strings.

Parameters:

mols (list[str | Mol]) – list of SMILES strings
smiles_standardizer (str, callable) – smiles standardizer to use
n_jobs (int) – number of parallel jobs to use
fill_value (float) – value to fill for missing features

Returns:

a tuple containing the QSPRDataset instance and a boolean mask indicating which molecules failed to be processed

Return type:

abstract fit(X: DataFrame | ndarray, y: DataFrame | ndarray, estimator: Any = None, mode: EarlyStoppingMode = EarlyStoppingMode.NOT_RECORDING, monitor: FitMonitor = None, **kwargs) → Any | tuple[Any, int] | None

Fit the model to the given data matrix or QSPRDataset.

Note. convertToNumpy can be called here, to convert the input data to: np.ndarray format.

Note. if no estimator is given, the estimator instance of the model is used.

Note. if a model supports early stopping, the fit function should have the: early_stopping decorator and the mode argument should be used to set the early stopping mode. If the model does not support early stopping, the mode argument is ignored.

Parameters:

X (pd.DataFrame, np.ndarray) – data matrix to fit
y (pd.DataFrame, np.ndarray) – target matrix to fit
estimator (Any) – estimator instance to use for fitting
mode (EarlyStoppingMode) – early stopping mode
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
kwargs – additional arguments to pass to the fit method of the estimator

Returns:

fitted estimator instance int: in case of early stopping, the number of iterations

after which the model stopped training

Return type:

Any

fitDataset(ds: QSPRDataset, monitor=None, mode=EarlyStoppingMode.OPTIMAL, save_model=True, save_data=False, **kwargs) → str

Train model on the whole attached data set.

** IMPORTANT ** For models that supportEarlyStopping, CrossValAssessor should be run first, so that the average number of epochs from the cross-validation with early stopping can be used for fitting the model.

Parameters:

ds (QSPRDataset) – data set to fit this model on
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
mode (EarlyStoppingMode) – early stopping mode for models that support early stopping, by default fit the ‘optimal’ number of epochs previously stopped at in model assessment on train or test set, to avoid the use of extra data for a validation set.
save_model (bool) – save the model to file
save_data (bool) – save the supplied dataset to file
kwargs – additional arguments to pass to fit

Returns:

path to the saved model, if save_model is True

Return type:

classmethod fromFile(filename: str) → QSPRModel

Initialize a new instance from a JSON file.

Parameters:: filename (str) – path to the JSON file
Returns:: new instance of the class
Return type:: instance (object)

classmethod fromJSON(json: str) → Any

Reconstruct object from a JSON string.

Parameters:: json (str) – JSON string of the object
Returns:: reconstructed object
Return type:: obj (object)

abstract getGPUs()[source]

getParameters(new_parameters) → dict | None

Get the model parameters combined with the given parameters.

If both the model and the given parameters contain the same key, the value from the given parameters is used.

Parameters:: new_parameters (dict) – dictionary of new parameters to add
Returns:: dictionary of model parameters
Return type:: dict

static handleInvalidsInPredictions(mols: list[str], predictions: ndarray | list[numpy.ndarray], failed_mask: ndarray) → ndarray

Replace invalid predictions with None.

Parameters:

mols (MoleculeTable) – molecules for which the predictions were made
predictions (np.ndarray) – predictions made by the model
failed_mask (np.ndarray) – boolean mask of failed predictions

Returns:

predictions with invalids replaced by None

Return type:

np.ndarray

initFromDataset(data: QSPRDataset | None)

initRandomState(random_state)

Set random state if applicable. Defaults to random state of dataset if no random state is provided,

Parameters:: random_state (int) – Random state to use for shuffling and other random operations.

property isMultiTask: bool

Return if model is a multitask model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: True if model is a multitask model
Return type:: bool

abstract loadEstimator(params: dict | None = None) → object

Initialize estimator instance with the given parameters.

If params is None, the default parameters will be used.

Parameters:: params (dict) – algorithm parameters
Returns:: initialized estimator instance
Return type:: object

abstract loadEstimatorFromFile(params: dict | None = None) → object

Load estimator instance from file and apply the given parameters.

Parameters:: params (dict) – algorithm parameters
Returns:: initialized estimator instance
Return type:: object

classmethod loadParamsGrid(fname: str, optim_type: str, model_types: str) → ndarray

Load parameter grids for bayes or grid search parameter optimization from json file.

Parameters:

fname (str) – file name of json file containing array with three columns containing modeltype, optimization type (grid or bayes) and model type
optim_type (str) – optimization type (grid or bayes)
model_types (list of str) – model type for hyperparameter optimization (e.g. RF)

Returns:

array with three columns containing modeltype, optimization type (grid or bayes) and model type

Return type:

np.ndarray

property metaFile: str

property optimalEpochs: int | None

Return the optimal number of epochs for early stopping.

Returns:: optimal number of epochs
Return type:: int | None

property outDir: str

Return output directory of the model, the model files are stored in this directory ({baseDir}/{name}).

Returns:: output directory of the model
Return type:: str

property outPrefix: str

Return output prefix of the model files.

The model files are stored with this prefix (i.e. {outPrefix}_meta.json).

Returns:: output prefix of the model files
Return type:: str

abstract predict(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → ndarray

Make predictions for the given data matrix or QSPRDataset.

Note. convertToNumpy can be called here, to convert the input data to np.ndarray format.

Note. if no estimator is given, the estimator instance of the model: is used.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to predict
estimator (Any) – estimator instance to use for fitting

Returns:

2D array containing the predictions, where each row corresponds to a sample in the data and each column to a target property

Return type:

np.ndarray

predictDataset(dataset: QSPRDataset, use_probas: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given dataset.

Parameters:

dataset – a QSPRDataset instance
use_probas – use probabilities if this is a classification model

Returns:

an array of predictions or a list of arrays of predictions (for classification models with use_probas=True)

Return type:

np.ndarray | list[np.ndarray]

predictMols(mols: List[str | Mol], use_probas: bool = False, smiles_standardizer: str | callable = 'chembl', n_jobs: int = 1, fill_value: float = nan, use_applicability_domain: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given molecules.

Parameters:

mols (List[str | Mol]) – list of SMILES strings
use_probas (bool) – use probabilities for classification models
smiles_standardizer – either chembl, old, or a partial function that reads and standardizes smiles.
n_jobs – Number of jobs to use for parallel processing.
fill_value – Value to use for missing values in the feature matrix.
use_applicability_domain – Use applicability domain to return if a molecule is within the applicability domain of the model.

Returns:

an array of predictions or a list of arrays of predictions: (for classification models with use_probas=True)
np.ndarray[bool]: boolean mask indicating which molecules fall: within the applicability domain of the model

Return type:

np.ndarray | list[np.ndarray]

abstract predictProba(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → list[numpy.ndarray]

Make predictions for the given data matrix or QSPRDataset, but use probabilities for classification models. Does not work with regression models.

Note. convertToNumpy can be called here, to convert the input data to np.ndarray format.

Note. if no estimator is given, the estimator instance of the model: is used.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to make predict
estimator (Any) – estimator instance to use for fitting

Returns:

a list of 2D arrays containing the probabilities for each class, where each array corresponds to a target property, each row to a sample in the data and each column to a class

Return type:

list[np.ndarray]

save(save_estimator=False)

Save model to file.

Parameters:

save_estimator (bool) – Explicitly save the estimator to file, if True. Note that some models may save the estimator by default even if this argument is False.

Returns:

absolute path to the metafile of the saved model str:

absolute path to the saved estimator, if include_estimator is True

Return type:

abstract saveEstimator() → str

Save the underlying estimator to file.

Returns:: absolute path to the saved estimator
Return type:: path (str)

abstract setGPUs(gpus: list[int])[source]

setParams(params: dict | None, reset_estimator: bool = True)

Set model parameters. The estimator is also updated with the new parameters if ‘reload_estimator’ is True.

Parameters:

params (dict) – dictionary of model parameters or None to reset the parameters
reset_estimator (bool) – if True, the estimator is reinitialized with the new parameters

abstract property supportsEarlyStopping: bool

Return if the model supports early stopping.

Returns:: True if the model supports early stopping
Return type:: bool

property task: ModelTasks

Return the task of the model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: task of the model
Return type:: ModelTasks

toFile(filename: str) → str

Serialize object to a JSON file. This JSON file should contain all data necessary to reconstruct the object.

Parameters:: filename (str) – filename to save object to
Returns:: absolute path to the saved JSON file of the object
Return type:: filename (str)

toJSON()

Serialize object to a JSON string. This JSON string should: contain all data necessary to reconstruct the object.

Returns:: JSON string of the object
Return type:: json (str)

qsprpred.extra.gpu.models.base_torch module

class qsprpred.extra.gpu.models.base_torch.QSPRModelPyTorchGPU(base_dir: str, alg: Type | None = None, name: str | None = None, parameters: dict | None = None, autoload=True, random_state: int | None = None)[source]

Bases: QSPRModelGPU, ABC

Initialize a QSPR model instance.

If the model is loaded from file, the data set is not required. Note that the data set is required for fitting and optimization.

Parameters:

base_dir (str) – base directory of the model, the model files are stored in a subdirectory {baseDir}/{outDir}/
alg (Type) – estimator class
name (str) – name of the model
parameters (dict) – dictionary of algorithm specific parameters
autoload (bool) – if True, the estimator is loaded from the serialized file if it exists, otherwise a new instance of alg is created
random_state (int) – Random state to use for shuffling and other random operations.

checkData(ds: QSPRDataset, exception: bool = True) → bool

Check if the model has a data set.

Parameters:

ds (QSPRDataset) – data set to check
exception (bool) – if true, an exception is raised if no data is set

Returns:

True if data is set, False otherwise (if exception is False)

Return type:

property classPath: str

Return the fully classified path of the model.

Returns:: class path of the model
Return type:: str

cleanFiles()

Clean up the model files.

Removes the model directory and all its contents.

Convert the given data matrix and target matrix to np.ndarray format.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix

Returns:

data matrix and/or target matrix in np.ndarray format

createPredictionDatasetFromMols(mols: list[str | rdkit.Chem.rdchem.Mol], smiles_standardizer: str | Callable[[str], str] = 'chembl', n_jobs: int = 1, fill_value: float = nan) → tuple[qsprpred.data.tables.qspr.QSPRDataset, numpy.ndarray]

Create a QSPRDataset instance from a list of SMILES strings.

Parameters:

mols (list[str | Mol]) – list of SMILES strings
smiles_standardizer (str, callable) – smiles standardizer to use
n_jobs (int) – number of parallel jobs to use
fill_value (float) – value to fill for missing features

Returns:

a tuple containing the QSPRDataset instance and a boolean mask indicating which molecules failed to be processed

Return type:

abstract fit(X: DataFrame | ndarray, y: DataFrame | ndarray, estimator: Any = None, mode: EarlyStoppingMode = EarlyStoppingMode.NOT_RECORDING, monitor: FitMonitor = None, **kwargs) → Any | tuple[Any, int] | None

Fit the model to the given data matrix or QSPRDataset.

Note. convertToNumpy can be called here, to convert the input data to: np.ndarray format.

Note. if no estimator is given, the estimator instance of the model is used.

Note. if a model supports early stopping, the fit function should have the: early_stopping decorator and the mode argument should be used to set the early stopping mode. If the model does not support early stopping, the mode argument is ignored.

Parameters:

X (pd.DataFrame, np.ndarray) – data matrix to fit
y (pd.DataFrame, np.ndarray) – target matrix to fit
estimator (Any) – estimator instance to use for fitting
mode (EarlyStoppingMode) – early stopping mode
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
kwargs – additional arguments to pass to the fit method of the estimator

Returns:

fitted estimator instance int: in case of early stopping, the number of iterations

after which the model stopped training

Return type:

Any

fitDataset(ds: QSPRDataset, monitor=None, mode=EarlyStoppingMode.OPTIMAL, save_model=True, save_data=False, **kwargs) → str

Train model on the whole attached data set.

** IMPORTANT ** For models that supportEarlyStopping, CrossValAssessor should be run first, so that the average number of epochs from the cross-validation with early stopping can be used for fitting the model.

Parameters:

ds (QSPRDataset) – data set to fit this model on
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
mode (EarlyStoppingMode) – early stopping mode for models that support early stopping, by default fit the ‘optimal’ number of epochs previously stopped at in model assessment on train or test set, to avoid the use of extra data for a validation set.
save_model (bool) – save the model to file
save_data (bool) – save the supplied dataset to file
kwargs – additional arguments to pass to fit

Returns:

path to the saved model, if save_model is True

Return type:

classmethod fromFile(filename: str) → QSPRModel

Initialize a new instance from a JSON file.

Parameters:: filename (str) – path to the JSON file
Returns:: new instance of the class
Return type:: instance (object)

classmethod fromJSON(json: str) → Any

Reconstruct object from a JSON string.

Parameters:: json (str) – JSON string of the object
Returns:: reconstructed object
Return type:: obj (object)

abstract getDevice() → device[source]

abstract getGPUs()[source]

getParameters(new_parameters) → dict | None

Get the model parameters combined with the given parameters.

If both the model and the given parameters contain the same key, the value from the given parameters is used.

Parameters:: new_parameters (dict) – dictionary of new parameters to add
Returns:: dictionary of model parameters
Return type:: dict

static handleInvalidsInPredictions(mols: list[str], predictions: ndarray | list[numpy.ndarray], failed_mask: ndarray) → ndarray

Replace invalid predictions with None.

Parameters:

mols (MoleculeTable) – molecules for which the predictions were made
predictions (np.ndarray) – predictions made by the model
failed_mask (np.ndarray) – boolean mask of failed predictions

Returns:

predictions with invalids replaced by None

Return type:

np.ndarray

initFromDataset(data: QSPRDataset | None)

initRandomState(random_state)

Set random state if applicable. Defaults to random state of dataset if no random state is provided,

Parameters:: random_state (int) – Random state to use for shuffling and other random operations.

property isMultiTask: bool

Return if model is a multitask model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: True if model is a multitask model
Return type:: bool

abstract loadEstimator(params: dict | None = None) → object

Initialize estimator instance with the given parameters.

If params is None, the default parameters will be used.

Parameters:: params (dict) – algorithm parameters
Returns:: initialized estimator instance
Return type:: object

abstract loadEstimatorFromFile(params: dict | None = None) → object

Load estimator instance from file and apply the given parameters.

Parameters:: params (dict) – algorithm parameters
Returns:: initialized estimator instance
Return type:: object

classmethod loadParamsGrid(fname: str, optim_type: str, model_types: str) → ndarray

Load parameter grids for bayes or grid search parameter optimization from json file.

Parameters:

fname (str) – file name of json file containing array with three columns containing modeltype, optimization type (grid or bayes) and model type
optim_type (str) – optimization type (grid or bayes)
model_types (list of str) – model type for hyperparameter optimization (e.g. RF)

Returns:

array with three columns containing modeltype, optimization type (grid or bayes) and model type

Return type:

np.ndarray

property metaFile: str

property optimalEpochs: int | None

Return the optimal number of epochs for early stopping.

Returns:: optimal number of epochs
Return type:: int | None

property outDir: str

Return output directory of the model, the model files are stored in this directory ({baseDir}/{name}).

Returns:: output directory of the model
Return type:: str

property outPrefix: str

Return output prefix of the model files.

The model files are stored with this prefix (i.e. {outPrefix}_meta.json).

Returns:: output prefix of the model files
Return type:: str

abstract predict(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → ndarray

Make predictions for the given data matrix or QSPRDataset.

Note. convertToNumpy can be called here, to convert the input data to np.ndarray format.

Note. if no estimator is given, the estimator instance of the model: is used.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to predict
estimator (Any) – estimator instance to use for fitting

Returns:

2D array containing the predictions, where each row corresponds to a sample in the data and each column to a target property

Return type:

np.ndarray

predictDataset(dataset: QSPRDataset, use_probas: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given dataset.

Parameters:

dataset – a QSPRDataset instance
use_probas – use probabilities if this is a classification model

Returns:

an array of predictions or a list of arrays of predictions (for classification models with use_probas=True)

Return type:

np.ndarray | list[np.ndarray]

predictMols(mols: List[str | Mol], use_probas: bool = False, smiles_standardizer: str | callable = 'chembl', n_jobs: int = 1, fill_value: float = nan, use_applicability_domain: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given molecules.

Parameters:

mols (List[str | Mol]) – list of SMILES strings
use_probas (bool) – use probabilities for classification models
smiles_standardizer – either chembl, old, or a partial function that reads and standardizes smiles.
n_jobs – Number of jobs to use for parallel processing.
fill_value – Value to use for missing values in the feature matrix.
use_applicability_domain – Use applicability domain to return if a molecule is within the applicability domain of the model.

Returns:

an array of predictions or a list of arrays of predictions: (for classification models with use_probas=True)
np.ndarray[bool]: boolean mask indicating which molecules fall: within the applicability domain of the model

Return type:

np.ndarray | list[np.ndarray]

abstract predictProba(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → list[numpy.ndarray]

Make predictions for the given data matrix or QSPRDataset, but use probabilities for classification models. Does not work with regression models.

Note. convertToNumpy can be called here, to convert the input data to np.ndarray format.

Note. if no estimator is given, the estimator instance of the model: is used.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to make predict
estimator (Any) – estimator instance to use for fitting

Returns:

a list of 2D arrays containing the probabilities for each class, where each array corresponds to a target property, each row to a sample in the data and each column to a class

Return type:

list[np.ndarray]

save(save_estimator=False)

Save model to file.

Parameters:

save_estimator (bool) – Explicitly save the estimator to file, if True. Note that some models may save the estimator by default even if this argument is False.

Returns:

absolute path to the metafile of the saved model str:

absolute path to the saved estimator, if include_estimator is True

Return type:

abstract saveEstimator() → str

Save the underlying estimator to file.

Returns:: absolute path to the saved estimator
Return type:: path (str)

abstract setDevice(device: str)[source]

abstract setGPUs(gpus: list[int])[source]

setParams(params: dict | None, reset_estimator: bool = True)

Set model parameters. The estimator is also updated with the new parameters if ‘reload_estimator’ is True.

Parameters:

params (dict) – dictionary of model parameters or None to reset the parameters
reset_estimator (bool) – if True, the estimator is reinitialized with the new parameters

abstract property supportsEarlyStopping: bool

Return if the model supports early stopping.

Returns:: True if the model supports early stopping
Return type:: bool

property task: ModelTasks

Return the task of the model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: task of the model
Return type:: ModelTasks

toFile(filename: str) → str

Serialize object to a JSON file. This JSON file should contain all data necessary to reconstruct the object.

Parameters:: filename (str) – filename to save object to
Returns:: absolute path to the saved JSON file of the object
Return type:: filename (str)

toJSON()

Serialize object to a JSON string. This JSON string should: contain all data necessary to reconstruct the object.

Returns:: JSON string of the object
Return type:: json (str)

qsprpred.extra.gpu.models.chemprop module

QSPRPpred wrapper for chemprop models.

class qsprpred.extra.gpu.models.chemprop.ChempropModel(base_dir: str, name: str | None = None, parameters: dict | None = None, autoload=True, random_state: int | None = None, quiet_logger: bool = True)[source]

Bases: QSPRModelPyTorchGPU

QSPRpred implementation of Chemprop model.

Variables:

name (str) – name of the model
alg (Type) – estimator class
parameters (dict) – dictionary of algorithm specific parameters
estimator (Any) – the underlying estimator instance of the type specified in QSPRModel.alg, if QSPRModel.fit or optimization was performed
featureCalculators (MoleculeDescriptorsCalculator) – feature calculator instance taken from the data set or deserialized from file if the model is loaded without data
featureStandardizer (SKLearnStandardizer) – feature standardizer instance taken from the data set or deserialized from file if the model is loaded without data
baseDir (str) – base directory of the model, the model files are stored in a subdirectory {baseDir}/{outDir}/

Initialize a Chemprop instance.

If the model is loaded from file, the data set is not required. Note that the data set is required for fitting and optimization.

Parameters:

base_dir (str) – base directory of the model, the model files are stored in a subdirectory {baseDir}/{outDir}/
name (str) – name of the model
parameters (dict) – dictionary of algorithm specific parameters
autoload (bool) – if True, the estimator is loaded from the serialized file if it exists, otherwise a new instance of alg is created
quiet_logger (bool) – if True, the chemprop logger is set to quiet mode (no debug messages)

checkArgs(args: TrainArgs | dict)[source]

Check if the given arguments are valid.

Parameters:: args (chemprop.args.TrainArgs, dict) – arguments to check

checkData(ds: QSPRDataset, exception: bool = True) → bool

Check if the model has a data set.

Parameters:

ds (QSPRDataset) – data set to check
exception (bool) – if true, an exception is raised if no data is set

Returns:

True if data is set, False otherwise (if exception is False)

Return type:

property classPath: str

Return the fully classified path of the model.

Returns:: class path of the model
Return type:: str

cleanFiles()[source]

Clean up the model files.

Removes the model directory and all its contents. Handles closing the chemprop logger as well.

Convert the given data matrix and target matrix to chemprop Molecule Dataset.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix

Returns:

data matrix and/or target matrix in np.ndarray format

Convert the given data matrix and target matrix to np.ndarray format.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix

Returns:

data matrix and/or target matrix in np.ndarray format

createPredictionDatasetFromMols(mols: list[str | rdkit.Chem.rdchem.Mol], smiles_standardizer: str | Callable[[str], str] = 'chembl', n_jobs: int = 1, fill_value: float = nan) → tuple[qsprpred.data.tables.qspr.QSPRDataset, numpy.ndarray]

Create a QSPRDataset instance from a list of SMILES strings.

Parameters:

mols (list[str | Mol]) – list of SMILES strings
smiles_standardizer (str, callable) – smiles standardizer to use
n_jobs (int) – number of parallel jobs to use
fill_value (float) – value to fill for missing features

Returns:

a tuple containing the QSPRDataset instance and a boolean mask indicating which molecules failed to be processed

Return type:

Wrapper for fit method of models that support early stopping.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to fit
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix to fit
estimator (Any) – estimator instance to use for fitting
mode (EarlyStoppingMode) – early stopping mode
split (DataSplit) – data split to use for early stopping, if None, a ShuffleSplit with 10% validation set size is used
monitor (FitMonitor) – monitor to use for fitting, if None, a BaseMonitor is used
kwargs (dict) – additional keyword arguments for the estimator’s fit method

Returns:

fitted estimator instance

Return type:

Any

fitDataset(ds: QSPRDataset, monitor=None, mode=EarlyStoppingMode.OPTIMAL, save_model=True, save_data=False, **kwargs) → str

Train model on the whole attached data set.

** IMPORTANT ** For models that supportEarlyStopping, CrossValAssessor should be run first, so that the average number of epochs from the cross-validation with early stopping can be used for fitting the model.

Parameters:

ds (QSPRDataset) – data set to fit this model on
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
mode (EarlyStoppingMode) – early stopping mode for models that support early stopping, by default fit the ‘optimal’ number of epochs previously stopped at in model assessment on train or test set, to avoid the use of extra data for a validation set.
save_model (bool) – save the model to file
save_data (bool) – save the supplied dataset to file
kwargs – additional arguments to pass to fit

Returns:

path to the saved model, if save_model is True

Return type:

classmethod fromFile(filename: str) → ChempropModel[source]

Initialize a new instance from a JSON file.

Parameters:: filename (str) – path to the JSON file
Returns:: new instance of the class
Return type:: instance (object)

classmethod fromJSON(json: str) → Any

Reconstruct object from a JSON string.

Parameters:: json (str) – JSON string of the object
Returns:: reconstructed object
Return type:: obj (object)

static getAvailableParameters()[source]

Return a dictionary of available parameters for the algorithm.

Definitions and default values can be found on the Chemprop github (https://github.com/chemprop/chemprop/blob/master/chemprop/args.py)

getDevice() → device[source]

getGPUs()[source]

getParameters(new_parameters) → dict | None

Get the model parameters combined with the given parameters.

If both the model and the given parameters contain the same key, the value from the given parameters is used.

Parameters:: new_parameters (dict) – dictionary of new parameters to add
Returns:: dictionary of model parameters
Return type:: dict

static handleInvalidsInPredictions(mols: list[str], predictions: ndarray | list[numpy.ndarray], failed_mask: ndarray) → ndarray

Replace invalid predictions with None.

Parameters:

mols (MoleculeTable) – molecules for which the predictions were made
predictions (np.ndarray) – predictions made by the model
failed_mask (np.ndarray) – boolean mask of failed predictions

Returns:

predictions with invalids replaced by None

Return type:

np.ndarray

initFromDataset(data: QSPRDataset | None)

initRandomState(random_state)

Set random state if applicable. Defaults to random state of dataset if no random state is provided,

Parameters:: random_state (int) – Random state to use for shuffling and other random operations.

property isMultiTask: bool

Return if model is a multitask model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: True if model is a multitask model
Return type:: bool

loadEstimator(params: dict | None = None) → object[source]

Initialize estimator instance with the given parameters.

If params is None, the default parameters will be used.

Parameters:: params (dict) – algorithm parameters
Returns:: initialized estimator instance
Return type:: object

loadEstimatorFromFile(params: dict | None = None, fallback_load=True) → object[source]

Load estimator instance from file and apply the given parameters.

Parameters:

params (dict) – algorithm parameters
fallback_load (bool) – if True, init estimator from alg if path not found

Returns:

initialized estimator instance

Return type:

object

classmethod loadParamsGrid(fname: str, optim_type: str, model_types: str) → ndarray

Load parameter grids for bayes or grid search parameter optimization from json file.

Parameters:

fname (str) – file name of json file containing array with three columns containing modeltype, optimization type (grid or bayes) and model type
optim_type (str) – optimization type (grid or bayes)
model_types (list of str) – model type for hyperparameter optimization (e.g. RF)

Returns:

array with three columns containing modeltype, optimization type (grid or bayes) and model type

Return type:

np.ndarray

property metaFile: str

property optimalEpochs: int | None

Return the optimal number of epochs for early stopping.

Returns:: optimal number of epochs
Return type:: int | None

property outDir: str

Return output directory of the model, the model files are stored in this directory ({baseDir}/{name}).

Returns:: output directory of the model
Return type:: str

property outPrefix: str

Return output prefix of the model files.

The model files are stored with this prefix (i.e. {outPrefix}_meta.json).

Returns:: output prefix of the model files
Return type:: str

predict(X: DataFrame | ndarray | QSPRDataset, estimator: ChempropMoleculeModel | None = None) → ndarray[source]

Make predictions for the given data matrix or QSPRDataset.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to predict
estimator (MoleculeModel) – estimator instance to use for fitting

Returns:

2D array containing the predictions, where each row corresponds to a sample in the data and each column to a target property

Return type:

np.ndarray

predictDataset(dataset: QSPRDataset, use_probas: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given dataset.

Parameters:

dataset – a QSPRDataset instance
use_probas – use probabilities if this is a classification model

Returns:

an array of predictions or a list of arrays of predictions (for classification models with use_probas=True)

Return type:

np.ndarray | list[np.ndarray]

predictMols(mols: List[str | Mol], use_probas: bool = False, smiles_standardizer: str | callable = 'chembl', n_jobs: int = 1, fill_value: float = nan, use_applicability_domain: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given molecules.

Parameters:

mols (List[str | Mol]) – list of SMILES strings
use_probas (bool) – use probabilities for classification models
smiles_standardizer – either chembl, old, or a partial function that reads and standardizes smiles.
n_jobs – Number of jobs to use for parallel processing.
fill_value – Value to use for missing values in the feature matrix.
use_applicability_domain – Use applicability domain to return if a molecule is within the applicability domain of the model.

Returns:

an array of predictions or a list of arrays of predictions: (for classification models with use_probas=True)
np.ndarray[bool]: boolean mask indicating which molecules fall: within the applicability domain of the model

Return type:

np.ndarray | list[np.ndarray]

predictProba(X: DataFrame | ndarray | QSPRDataset, estimator: ChempropMoleculeModel | None = None) → list[numpy.ndarray][source]

Make predictions for the given data matrix or QSPRDataset, but use probabilities for classification models.

In case of regression models, this method is equivalent to predict.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to make predict
estimator (MoleculeModel, None) – estimator instance to use for fitting

Returns:

a list of 2D arrays containing the probabilities for each class, where each array corresponds to a target property, each row to a sample in the data and each column to a class

Return type:

list[np.ndarray]

save(save_estimator=False)

Save model to file.

Parameters:

save_estimator (bool) – Explicitly save the estimator to file, if True. Note that some models may save the estimator by default even if this argument is False.

Returns:

absolute path to the metafile of the saved model str:

absolute path to the saved estimator, if include_estimator is True

Return type:

saveEstimator() → str[source]

Save the underlying estimator to file.

Returns:: path to the saved estimator
Return type:: path (str)

setDevice(device: str)[source]

setGPUs(gpus: list[int])[source]

setParams(params: dict | None, reset_estimator: bool = True)

Set model parameters. The estimator is also updated with the new parameters if ‘reload_estimator’ is True.

Parameters:

params (dict) – dictionary of model parameters or None to reset the parameters
reset_estimator (bool) – if True, the estimator is reinitialized with the new parameters

supportsEarlyStopping() → bool[source]

Return if the model supports early stopping.

Returns:: True if the model supports early stopping
Return type:: bool

property task: ModelTasks

Return the task of the model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: task of the model
Return type:: ModelTasks

toFile(filename: str) → str

Serialize object to a JSON file. This JSON file should contain all data necessary to reconstruct the object.

Parameters:: filename (str) – filename to save object to
Returns:: absolute path to the saved JSON file of the object
Return type:: filename (str)

toJSON()

Serialize object to a JSON string. This JSON string should: contain all data necessary to reconstruct the object.

Returns:: JSON string of the object
Return type:: json (str)

class qsprpred.extra.gpu.models.chemprop.ChempropMoleculeModel(args: TrainArgs, scaler: StandardScaler | None = None)[source]

Bases: MoleculeModel

Wrapper for chemprop.models.MoleculeModel.

Variables:

args (chemprop.args.TrainArgs) – arguments for training the model,
scaler (chemprop.data.scaler.StandardScaler) – scaler for scaling the targets

Initialize a MoleculeModel instance.

Parameters:

args (chemprop.args.TrainArgs) – arguments for training the model,
scaler (chemprop.data.scaler.StandardScaler) – scaler for scaling the targets

T_destination = ~T_destination

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

name (str) – name of the child module. The child module can be accessed from this module using the given name
module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:: fn (Module -> None) – function to be applied to each submodule
Returns:: self
Return type:: Module

Example:

>>> @torch.no_grad()
>>> def init_weights(m):
>>>     print(m)
>>>     if type(m) == nn.Linear:
>>>         m.weight.fill_(1.0)
>>>         print(m.weight)
>>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2))
>>> net.apply(init_weights)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:: recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.
Yields:: torch.Tensor – module buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for buf in model.buffers():
>>>     print(type(buf), buf.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

call_super_init: bool = False

classmethod cast(obj: MoleculeModel) → ChempropMoleculeModel[source]

Cast a chemprop.models.MoleculeModel instance to a MoleculeModel instance.

Parameters:: obj (chemprop.models.MoleculeModel) – instance to cast
Returns:: casted MoleculeModel instance
Return type:: MoleculeModel

children() → Iterator[Module]

Return an iterator over immediate children modules.

Yields:: Module – a child module

compile(*args, **kwargs)

Compile this Module’s forward using torch.compile().

This Module’s __call__ method is compiled and all arguments are passed as-is to torch.compile().

See torch.compile() for details on the arguments for this function.

cpu() → T

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

create_encoder(args: TrainArgs) → None

Creates the message passing encoder for the model.

Parameters:: args – A TrainArgs object containing model arguments.

create_ffn(args: TrainArgs) → None

Creates the feed-forward layers for the model.

Parameters:: args – A TrainArgs object containing model arguments.

cuda(device: int | device | None = None) → T

Move all model parameters and buffers to the GPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on GPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

dump_patches: bool = False

eval() → T

Set the module in evaluation mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

This is equivalent with self.train(False).

See locally-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:: self
Return type:: Module

extra_repr() → str

Set the extra representation of the module.

To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.

fingerprint(batch: List[List[str]] | List[List[Mol]] | List[List[Tuple[Mol, Mol]]] | List[BatchMolGraph], features_batch: List[ndarray] = None, atom_descriptors_batch: List[ndarray] = None, atom_features_batch: List[ndarray] = None, bond_descriptors_batch: List[ndarray] = None, bond_features_batch: List[ndarray] = None, fingerprint_type: str = 'MPN') → Tensor

Encodes the latent representations of the input molecules from intermediate stages of the model.

Parameters:

batch – A list of list of SMILES, a list of list of RDKit molecules, or a list of BatchMolGraph. The outer list or BatchMolGraph is of length num_molecules (number of datapoints in batch), the inner list is of length number_of_molecules (number of molecules per datapoint).
features_batch – A list of numpy arrays containing additional features.
atom_descriptors_batch – A list of numpy arrays containing additional atom descriptors.
atom_features_batch – A list of numpy arrays containing additional atom features.
bond_descriptors_batch – A list of numpy arrays containing additional bond descriptors.
bond_features_batch – A list of numpy arrays containing additional bond features.
fingerprint_type – The choice of which type of latent representation to return as the molecular fingerprint. Currently supported MPN for the output of the MPNN portion of the model or last_FFN for the input to the final readout layer.

Returns:

The latent fingerprint vectors.

float() → T

Casts all floating point parameters and buffers to float datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

forward(batch: List[List[str]] | List[List[Mol]] | List[List[Tuple[Mol, Mol]]] | List[BatchMolGraph], features_batch: List[ndarray] = None, atom_descriptors_batch: List[ndarray] = None, atom_features_batch: List[ndarray] = None, bond_descriptors_batch: List[ndarray] = None, bond_features_batch: List[ndarray] = None, constraints_batch: List[Tensor] = None, bond_types_batch: List[Tensor] = None) → Tensor

Runs the MoleculeModel on input.

Parameters:

batch – A list of list of SMILES, a list of list of RDKit molecules, or a list of BatchMolGraph. The outer list or BatchMolGraph is of length num_molecules (number of datapoints in batch), the inner list is of length number_of_molecules (number of molecules per datapoint).
features_batch – A list of numpy arrays containing additional features.
atom_descriptors_batch – A list of numpy arrays containing additional atom descriptors.
atom_features_batch – A list of numpy arrays containing additional atom features.
bond_descriptors_batch – A list of numpy arrays containing additional bond descriptors.
bond_features_batch – A list of numpy arrays containing additional bond features.
constraints_batch – A list of PyTorch tensors which applies constraint on atomic/bond properties.
bond_types_batch – A list of PyTorch tensors storing bond types of each bond determined by RDKit molecules.

Returns:

The output of the MoleculeModel, containing a list of property predictions.

static getTrainArgs(args: dict | None, task: ModelTasks) → TrainArgs[source]

Get a chemprop.args.TrainArgs instance from a dictionary.

Parameters:

args (dict) – dictionary of arguments
task (ModelTasks) – task type

Returns:

arguments for training the model

Return type:

chemprop.args.TrainArgs

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The buffer referenced by target
Return type:: torch.Tensor
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return any extra state to include in the module’s state_dict.

Implement this and a corresponding set_extra_state() for your module if you need to store extra state. This function is called when building the module’s state_dict().

Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes.

Returns:: Any extra state to store in the module’s state_dict
Return type:: object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The Parameter referenced by target
Return type:: torch.nn.Parameter
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw an error.

For example, let’s say you have an nn.Module A that looks like this:

A(
    (net_b): Module(
        (net_c): Module(
            (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2))
        )
        (linear): Linear(in_features=100, out_features=200, bias=True)
    )
)

(The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used.

Parameters:: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)
Returns:: The submodule referenced by target
Return type:: torch.nn.Module
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

half() → T

Casts all floating point parameters and buffers to half datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

ipu(device: int | device | None = None) → T

Move all model parameters and buffers to the IPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on IPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

state_dict (dict) – a dict containing parameters and persistent buffers.
strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True
assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

missing_keys is a list of str containing any keys that are expected
by this module but missing from the provided state_dict.
unexpected_keys is a list of str containing the keys that are not
expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:: Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.modules()):
...     print(idx, '->', m)

0 -> Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

prefix (str) – prefix to prepend to all buffer names.
recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True.
remove_duplicate (bool, optional) – whether to remove the duplicated buffers in the result. Defaults to True.

Yields:

(str, torch.Tensor) – Tuple containing the name and buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:: (str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

memo – a memo to store the set of modules already added to the result
prefix – a prefix that will be added to the name of the module
remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.named_modules()):
...     print(idx, '->', m)

0 -> ('', Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

prefix (str) – prefix to prepend to all parameter names.
recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:: recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
Yields:: Parameter – module parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for param in model.parameters():
>>>     print(type(param), param.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register a backward hook on the module.

This function is deprecated in favor of register_full_backward_hook() and the behavior of this function will change in future versions.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

name (str) – name of the buffer. The buffer can be accessed from this module using the given name
tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.
persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature:

hook(module, args, kwargs, output) -> None or modified output

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If True, the provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If True, the hook will be passed the kwargs given to the forward function. Default: False
always_call (bool) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_forward_pre_hook(hook: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register a forward pre-hook on the module.

The hook will be called every time before forward() is invoked.

If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature:

hook(module, args) -> None or modified input

If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature:

hook(module, args, kwargs) -> None or a tuple of modified input and kwargs

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If true, the hook will be passed the kwargs given to the forward function. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward hook on the module.

The hook will be called every time the gradients with respect to a module are computed, i.e. the hook will execute if and only if the gradients with respect to module outputs are computed. The hook should have the following signature:

hook(module, grad_input, grad_output) -> tuple(Tensor) or None

The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_pre_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward pre-hook on the module.

The hook will be called every time the gradients for the module are computed. The hook should have the following signature:

hook(module, grad_output) -> tuple[Tensor] or None

The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_load_state_dict_post_hook(hook)

Register a post hook to be run after module’s load_state_dict is called.

It should have the following signature::: hook(module, incompatible_keys) -> None

The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys.

The given incompatible_keys can be modified inplace if needed.

Note that the checks performed when calling load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_module(name: str, module: Module | None) → None: Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

name (str) – name of the parameter. The parameter can be accessed from this module using the given name
param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

These hooks will be called with arguments: self, prefix, and keep_vars before calling state_dict on self. The registered hooks can be used to perform pre-processing before the state_dict call is made.

requires_grad_(requires_grad: bool = True) → T

Change if autograd should record operations on parameters in this module.

This method sets the parameters’ requires_grad attributes in-place.

This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training).

See locally-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:: requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.
Returns:: self
Return type:: Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_extra_state() for your module if you need to store extra state within its state_dict.

Parameters:: state (dict) – Extra state from the state_dict

share_memory() → T: See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included.

Note

The returned object is a shallow copy. It contains references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

destination (dict, optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None.
prefix (str, optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''.
keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast the parameters and buffers.

This can be called as

to(device=None, dtype=None, non_blocking=False)

to(dtype, non_blocking=False)

to(tensor, non_blocking=False)

to(memory_format=torch.channels_last)

Its signature is similar to torch.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices.

See below for examples.

Note

This method modifies the module in-place.

Parameters:

device (torch.device) – the desired device of the parameters and buffers in this module
dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module
tensor (torch.Tensor) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module
memory_format (torch.memory_format) – the desired memory format for 4D parameters and buffers in this module (keyword only argument)

Returns:

self

Return type:

Module

Examples:

>>> # xdoctest: +IGNORE_WANT("non-deterministic")
>>> linear = nn.Linear(2, 2)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]])
>>> linear.to(torch.double)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]], dtype=torch.float64)
>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_CUDA1)
>>> gpu1 = torch.device("cuda:1")
>>> linear.to(gpu1, dtype=torch.half, non_blocking=True)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16, device='cuda:1')
>>> cpu = torch.device("cpu")
>>> linear.to(cpu)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16)

>>> linear = nn.Linear(2, 2, bias=None).to(torch.cdouble)
>>> linear.weight
Parameter containing:
tensor([[ 0.3741+0.j,  0.2382+0.j],
        [ 0.5593+0.j, -0.4443+0.j]], dtype=torch.complex128)
>>> linear(torch.ones(3, 2, dtype=torch.cdouble))
tensor([[0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j]], dtype=torch.complex128)

to_empty(*, device: int | str | device | None, recurse: bool = True) → T

Move the parameters and buffers to the specified device without copying storage.

Parameters:

device (torch.device) – The desired device of the parameters and buffers in this module.
recurse (bool) – Whether parameters and buffers of submodules should be recursively moved to the specified device.

Returns:

self

Return type:

Module

train(mode: bool = True) → T

Set the module in training mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

Parameters:: mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.
Returns:: self
Return type:: Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:: dst_type (type or string) – the desired type
Returns:: self
Return type:: Module

xpu(device: int | device | None = None) → T

Move all model parameters and buffers to the XPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on XPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

zero_grad(set_to_none: bool = True) → None

Reset gradients of all model parameters.

See similar function under torch.optim.Optimizer for more context.

Parameters:: set_to_none (bool) – instead of setting to zero, set the grads to None. See torch.optim.Optimizer.zero_grad() for details.

qsprpred.extra.gpu.models.dnn module

Here the DNN model originally from DrugEx can be found.

At the moment this contains a class for fully-connected DNNs.

class qsprpred.extra.gpu.models.dnn.DNNModel(base_dir: str, alg: ~typing.Type = <class 'qsprpred.extra.gpu.models.neural_network.STFullyConnected'>, name: str | None = None, parameters: dict | None = None, random_state: int | None = None, autoload: bool = True, gpus: list[int] = (0, ), patience: int = 50, tol: float = 0)[source]

Bases: QSPRModelPyTorchGPU

This class holds the methods for training and fitting a Deep Neural Net QSPR model initialization.

Here the model instance is created and parameters can be defined.

Variables:

name (str) – name of the model
alg (estimator) – estimator instance or class
parameters (dict) – dictionary of algorithm specific parameters
estimator (object) – the underlying estimator instance, if fit or optimization is performed, this model instance gets updated accordingly
featureCalculators (MoleculeDescriptorsCalculator) – feature calculator instance taken from the data set or deserialized from file if the model is loaded without data
featureStandardizer (SKLearnStandardizer) – feature standardizer instance taken from the data set or deserialized from file if the model is loaded without data
baseDir (str) – base directory of the model, the model files are stored in a subdirectory {baseDir}/{outDir}/
patience (int) – number of epochs to wait before early stop if no progress on validation set score
tol (float) – minimum absolute improvement of loss necessary to count as progress on best validation score
nClass (int) – number of classes
nDim (int) – number of features
patience – number of epochs to wait before early stop if no progress on validation set score

Initialize a DNNModel model.

Parameters:

base_dir (str) – base directory of the model, the model files are stored in a subdirectory {baseDir}/{outDir}/
alg (Type, optional) – model class or instance. Defaults to STFullyConnected.
name (str, optional) – name of the model. Defaults to None.
parameters (dict, optional) – dictionary of algorithm specific parameters. Defaults to None.
autoload (bool, optional) – whether to load the model from file or not. Defaults to True.
device (torch.device, optional) – The cuda device. Defaults to DEFAULT_TORCH_DEVICE.
gpus (list[int], optional) – gpu number(s) to use for model fitting. Defaults to DEFAULT_TORCH_GPUS.
patience (int, optional) – number of epochs to wait before early stop if no progress on validation set score. Defaults to 50.
tol (float, optional) – minimum absolute improvement of loss necessary to count as progress on best validation score. Defaults to 0.

checkData(ds: QSPRDataset, exception: bool = True) → bool

Check if the model has a data set.

Parameters:

ds (QSPRDataset) – data set to check
exception (bool) – if true, an exception is raised if no data is set

Returns:

True if data is set, False otherwise (if exception is False)

Return type:

property classPath: str

Return the fully classified path of the model.

Returns:: class path of the model
Return type:: str

cleanFiles()

Clean up the model files.

Removes the model directory and all its contents.

Convert the given data matrix and target matrix to np.ndarray format.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix

Returns:

data matrix and/or target matrix in np.ndarray format

createPredictionDatasetFromMols(mols: list[str | rdkit.Chem.rdchem.Mol], smiles_standardizer: str | Callable[[str], str] = 'chembl', n_jobs: int = 1, fill_value: float = nan) → tuple[qsprpred.data.tables.qspr.QSPRDataset, numpy.ndarray]

Create a QSPRDataset instance from a list of SMILES strings.

Parameters:

mols (list[str | Mol]) – list of SMILES strings
smiles_standardizer (str, callable) – smiles standardizer to use
n_jobs (int) – number of parallel jobs to use
fill_value (float) – value to fill for missing features

Returns:

a tuple containing the QSPRDataset instance and a boolean mask indicating which molecules failed to be processed

Return type:

Wrapper for fit method of models that support early stopping.

Parameters:

X (pd.DataFrame, np.ndarray, QSPRDataset) – data matrix to fit
y (pd.DataFrame, np.ndarray, QSPRDataset) – target matrix to fit
estimator (Any) – estimator instance to use for fitting
mode (EarlyStoppingMode) – early stopping mode
split (DataSplit) – data split to use for early stopping, if None, a ShuffleSplit with 10% validation set size is used
monitor (FitMonitor) – monitor to use for fitting, if None, a BaseMonitor is used
kwargs (dict) – additional keyword arguments for the estimator’s fit method

Returns:

fitted estimator instance

Return type:

Any

fitDataset(ds: QSPRDataset, monitor=None, mode=EarlyStoppingMode.OPTIMAL, save_model=True, save_data=False, **kwargs) → str

Train model on the whole attached data set.

** IMPORTANT ** For models that supportEarlyStopping, CrossValAssessor should be run first, so that the average number of epochs from the cross-validation with early stopping can be used for fitting the model.

Parameters:

ds (QSPRDataset) – data set to fit this model on
monitor (FitMonitor) – monitor for the fitting process, if None, the base monitor is used
mode (EarlyStoppingMode) – early stopping mode for models that support early stopping, by default fit the ‘optimal’ number of epochs previously stopped at in model assessment on train or test set, to avoid the use of extra data for a validation set.
save_model (bool) – save the model to file
save_data (bool) – save the supplied dataset to file
kwargs – additional arguments to pass to fit

Returns:

path to the saved model, if save_model is True

Return type:

classmethod fromFile(filename: str) → QSPRModel

Initialize a new instance from a JSON file.

Parameters:: filename (str) – path to the JSON file
Returns:: new instance of the class
Return type:: instance (object)

classmethod fromJSON(json: str) → Any

Reconstruct object from a JSON string.

Parameters:: json (str) – JSON string of the object
Returns:: reconstructed object
Return type:: obj (object)

getDevice() → device[source]

getGPUs()[source]

getParameters(new_parameters) → dict | None

Get the model parameters combined with the given parameters.

If both the model and the given parameters contain the same key, the value from the given parameters is used.

Parameters:: new_parameters (dict) – dictionary of new parameters to add
Returns:: dictionary of model parameters
Return type:: dict

static handleInvalidsInPredictions(mols: list[str], predictions: ndarray | list[numpy.ndarray], failed_mask: ndarray) → ndarray

Replace invalid predictions with None.

Parameters:

mols (MoleculeTable) – molecules for which the predictions were made
predictions (np.ndarray) – predictions made by the model
failed_mask (np.ndarray) – boolean mask of failed predictions

Returns:

predictions with invalids replaced by None

Return type:

np.ndarray

initFromDataset(data: QSPRDataset | None)[source]

initRandomState(random_state)[source]

Set random state if applicable. Defaults to random state of dataset if no random state is provided by the constructor.

Parameters:: random_state (int) – Random state to use for shuffling and other random operations.

property isMultiTask: bool

Return if model is a multitask model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: True if model is a multitask model
Return type:: bool

loadEstimator(params: dict | None = None) → object[source]

Load model from file or initialize new model.

Parameters:: params (dict, optional) – model parameters. Defaults to None.
Returns:: model instance
Return type:: model (object)

loadEstimatorFromFile(params: dict | None = None, fallback_load: bool = True) → object[source]

Load estimator from file.

Parameters:

params (dict) – parameters
fallback_load (bool) – if True, init estimator from alg and params if no estimator found at path

Returns:

estimator instance

Return type:

estimator (object)

classmethod loadParamsGrid(fname: str, optim_type: str, model_types: str) → ndarray

Load parameter grids for bayes or grid search parameter optimization from json file.

Parameters:

fname (str) – file name of json file containing array with three columns containing modeltype, optimization type (grid or bayes) and model type
optim_type (str) – optimization type (grid or bayes)
model_types (list of str) – model type for hyperparameter optimization (e.g. RF)

Returns:

array with three columns containing modeltype, optimization type (grid or bayes) and model type

Return type:

np.ndarray

property metaFile: str

property optimalEpochs: int | None

Return the optimal number of epochs for early stopping.

Returns:: optimal number of epochs
Return type:: int | None

property outDir: str

Return output directory of the model, the model files are stored in this directory ({baseDir}/{name}).

Returns:: output directory of the model
Return type:: str

property outPrefix: str

Return output prefix of the model files.

The model files are stored with this prefix (i.e. {outPrefix}_meta.json).

Returns:: output prefix of the model files
Return type:: str

predict(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → ndarray[source]: See QSPRModel.predict.

predictDataset(dataset: QSPRDataset, use_probas: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given dataset.

Parameters:

dataset – a QSPRDataset instance
use_probas – use probabilities if this is a classification model

Returns:

an array of predictions or a list of arrays of predictions (for classification models with use_probas=True)

Return type:

np.ndarray | list[np.ndarray]

predictMols(mols: List[str | Mol], use_probas: bool = False, smiles_standardizer: str | callable = 'chembl', n_jobs: int = 1, fill_value: float = nan, use_applicability_domain: bool = False) → ndarray | list[numpy.ndarray]

Make predictions for the given molecules.

Parameters:

mols (List[str | Mol]) – list of SMILES strings
use_probas (bool) – use probabilities for classification models
smiles_standardizer – either chembl, old, or a partial function that reads and standardizes smiles.
n_jobs – Number of jobs to use for parallel processing.
fill_value – Value to use for missing values in the feature matrix.
use_applicability_domain – Use applicability domain to return if a molecule is within the applicability domain of the model.

Returns:

an array of predictions or a list of arrays of predictions: (for classification models with use_probas=True)
np.ndarray[bool]: boolean mask indicating which molecules fall: within the applicability domain of the model

Return type:

np.ndarray | list[np.ndarray]

predictProba(X: DataFrame | ndarray | QSPRDataset, estimator: Any = None) → ndarray[source]: See QSPRModel.predictProba.

save(save_estimator=False)

Save model to file.

Parameters:

save_estimator (bool) – Explicitly save the estimator to file, if True. Note that some models may save the estimator by default even if this argument is False.

Returns:

absolute path to the metafile of the saved model str:

absolute path to the saved estimator, if include_estimator is True

Return type:

saveEstimator() → str[source]

Save the DNNModel model.

Returns:: path to the saved model
Return type:: str

setDevice(device: str)[source]

setGPUs(gpus: list[int])[source]

setParams(params: dict | None, reset_estimator: bool = True)

Set model parameters. The estimator is also updated with the new parameters if ‘reload_estimator’ is True.

Parameters:

params (dict) – dictionary of model parameters or None to reset the parameters
reset_estimator (bool) – if True, the estimator is reinitialized with the new parameters

property supportsEarlyStopping: bool: Whether the model supports early stopping or not.

property task: ModelTasks

Return the task of the model, taken from the data set or deserialized from file if the model is loaded without data.

Returns:: task of the model
Return type:: ModelTasks

toFile(filename: str) → str

Serialize object to a JSON file. This JSON file should contain all data necessary to reconstruct the object.

Parameters:: filename (str) – filename to save object to
Returns:: absolute path to the saved JSON file of the object
Return type:: filename (str)

toJSON()

Serialize object to a JSON string. This JSON string should: contain all data necessary to reconstruct the object.

Returns:: JSON string of the object
Return type:: json (str)

qsprpred.extra.gpu.models.neural_network module

This module holds the base class for DNN models as well as fully connected NN subclass.

class qsprpred.extra.gpu.models.neural_network.Base(device: str, gpus: list[int], n_epochs: int = 1000, lr: float = 0.0001, batch_size: int = 256, patience: int = 50, tol: float = 0)[source]

Bases: Module

Base structure for all classification/regression DNN models.

Mainly, it provides the general methods for training, evaluating model and predicting the given data.

Variables:

n_epochs (int) – (maximum) number of epochs to train the model
lr (float) – learning rate
batch_size (int) – batch size
patience (int) – number of epochs to wait before early stop if no progress on validation set score, if patience = -1, always train to n_epochs
tol (float) – minimum absolute improvement of loss necessary to count as progress on best validation score
device (torch.device) – device to run the model on
gpus (list) – list of gpus to run the model on

Initialize the DNN model.

Parameters:

device (str) – device to run the model on
gpus (list) – list of gpus to run the model on
n_epochs (int) – (maximum) number of epochs to train the model
lr (float) – learning rate
batch_size (int) – batch size
patience (int) – number of epochs to wait before early stop if no progress on validation set score, if patience = -1, always train to n_epochs
tol (float) – minimum absolute improvement of loss necessary to count as progress on best validation score

T_destination = ~T_destination

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

name (str) – name of the child module. The child module can be accessed from this module using the given name
module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:: fn (Module -> None) – function to be applied to each submodule
Returns:: self
Return type:: Module

Example:

>>> @torch.no_grad()
>>> def init_weights(m):
>>>     print(m)
>>>     if type(m) == nn.Linear:
>>>         m.weight.fill_(1.0)
>>>         print(m.weight)
>>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2))
>>> net.apply(init_weights)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:: recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.
Yields:: torch.Tensor – module buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for buf in model.buffers():
>>>     print(type(buf), buf.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

call_super_init: bool = False

children() → Iterator[Module]

Return an iterator over immediate children modules.

Yields:: Module – a child module

compile(*args, **kwargs)

Compile this Module’s forward using torch.compile().

This Module’s __call__ method is compiled and all arguments are passed as-is to torch.compile().

See torch.compile() for details on the arguments for this function.

cpu() → T

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

cuda(device: int | device | None = None) → T

Move all model parameters and buffers to the GPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on GPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

dump_patches: bool = False

eval() → T

Set the module in evaluation mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

This is equivalent with self.train(False).

See locally-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:: self
Return type:: Module

evaluate(loader) → float[source]

Evaluate the performance of the DNN model.

Parameters:: loader (torch.util.data.DataLoader) – data loader for test set, including m X n target FloatTensor and l X n label FloatTensor (m is the No. of sample, n is the No. of features, l is the No. of classes or tasks)
Returns:: the average loss value based on the calculation of loss function with given test set.
Return type:: loss (float)

extra_repr() → str

Set the extra representation of the module.

To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.

fit(X_train, y_train, X_valid=None, y_valid=None, monitor: FitMonitor | None = None) → int[source]

Training the DNN model.

Training is, similar to the scikit-learn or Keras style. It saves the optimal value of parameters.

Parameters:

X_train (np.ndarray or pd.Dataframe) – training data (m X n), m is the No. of samples, n is the No. of features
y_train (np.ndarray or pd.Dataframe) – training target (m X l), m is the No. of samples, l is the No. of classes or tasks
X_valid (np.ndarray or pd.Dataframe) – validation data (m X n), m is the No. of samples, n is the No. of features
y_valid (np.ndarray or pd.Dataframe) – validation target (m X l), m is the No. of samples, l is the No. of classes or tasks
monitor (FitMonitor) – monitor to use for training, if None, use base monitor

Returns:

the epoch number when the optimal model is saved

Return type:

int

float() → T

Casts all floating point parameters and buffers to float datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

forward(*input: Any) → None

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

getDataLoader(X, y=None)[source]

Convert data to tensors and get generator over dataset with dataloader.

Parameters:

X (numpy 2d array) – input dataset
y (numpy 1d column vector) – output data

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The buffer referenced by target
Return type:: torch.Tensor
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return any extra state to include in the module’s state_dict.

Implement this and a corresponding set_extra_state() for your module if you need to store extra state. This function is called when building the module’s state_dict().

Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes.

Returns:: Any extra state to store in the module’s state_dict
Return type:: object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The Parameter referenced by target
Return type:: torch.nn.Parameter
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_params(deep=True) → dict[source]

Get parameters for this estimator.

Function copied from sklearn.base_estimator!

Parameters:: deep (bool) – If True, will return the parameters for this estimator
Returns:: Parameter names mapped to their values.
Return type:: params (dict)

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw an error.

For example, let’s say you have an nn.Module A that looks like this:

A(
    (net_b): Module(
        (net_c): Module(
            (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2))
        )
        (linear): Linear(in_features=100, out_features=200, bias=True)
    )
)

(The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used.

Parameters:: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)
Returns:: The submodule referenced by target
Return type:: torch.nn.Module
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

half() → T

Casts all floating point parameters and buffers to half datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

ipu(device: int | device | None = None) → T

Move all model parameters and buffers to the IPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on IPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

state_dict (dict) – a dict containing parameters and persistent buffers.
strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True
assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

missing_keys is a list of str containing any keys that are expected
by this module but missing from the provided state_dict.
unexpected_keys is a list of str containing the keys that are not
expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:: Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.modules()):
...     print(idx, '->', m)

0 -> Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

prefix (str) – prefix to prepend to all buffer names.
recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True.
remove_duplicate (bool, optional) – whether to remove the duplicated buffers in the result. Defaults to True.

Yields:

(str, torch.Tensor) – Tuple containing the name and buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:: (str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

memo – a memo to store the set of modules already added to the result
prefix – a prefix that will be added to the name of the module
remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.named_modules()):
...     print(idx, '->', m)

0 -> ('', Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

prefix (str) – prefix to prepend to all parameter names.
recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:: recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
Yields:: Parameter – module parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for param in model.parameters():
>>>     print(type(param), param.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

predict(X_test) → ndarray[source]

Predicting the probability of each sample in the given dataset.

Parameters:: X_test (ndarray) – m X n target array (m is the No. of sample, n is the No. of features)
Returns:: probability of each sample in the given dataset, it is an m X l FloatTensor (m is the No. of sample, l is the No. of classes or tasks.)
Return type:: score (ndarray)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register a backward hook on the module.

This function is deprecated in favor of register_full_backward_hook() and the behavior of this function will change in future versions.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

name (str) – name of the buffer. The buffer can be accessed from this module using the given name
tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.
persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature:

hook(module, args, kwargs, output) -> None or modified output

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If True, the provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If True, the hook will be passed the kwargs given to the forward function. Default: False
always_call (bool) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_forward_pre_hook(hook: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register a forward pre-hook on the module.

The hook will be called every time before forward() is invoked.

If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature:

hook(module, args) -> None or modified input

If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature:

hook(module, args, kwargs) -> None or a tuple of modified input and kwargs

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If true, the hook will be passed the kwargs given to the forward function. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward hook on the module.

The hook will be called every time the gradients with respect to a module are computed, i.e. the hook will execute if and only if the gradients with respect to module outputs are computed. The hook should have the following signature:

hook(module, grad_input, grad_output) -> tuple(Tensor) or None

The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_pre_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward pre-hook on the module.

The hook will be called every time the gradients for the module are computed. The hook should have the following signature:

hook(module, grad_output) -> tuple[Tensor] or None

The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_load_state_dict_post_hook(hook)

Register a post hook to be run after module’s load_state_dict is called.

It should have the following signature::: hook(module, incompatible_keys) -> None

The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys.

The given incompatible_keys can be modified inplace if needed.

Note that the checks performed when calling load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_module(name: str, module: Module | None) → None: Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

name (str) – name of the parameter. The parameter can be accessed from this module using the given name
param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

These hooks will be called with arguments: self, prefix, and keep_vars before calling state_dict on self. The registered hooks can be used to perform pre-processing before the state_dict call is made.

requires_grad_(requires_grad: bool = True) → T

Change if autograd should record operations on parameters in this module.

This method sets the parameters’ requires_grad attributes in-place.

This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training).

See locally-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:: requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.
Returns:: self
Return type:: Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_extra_state() for your module if you need to store extra state within its state_dict.

Parameters:: state (dict) – Extra state from the state_dict

set_params(**params) → Base[source]

Set the parameters of this estimator.

Function copied from sklearn.base_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:: estimator instance
Return type:: self

share_memory() → T: See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included.

Note

The returned object is a shallow copy. It contains references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

destination (dict, optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None.
prefix (str, optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''.
keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast the parameters and buffers.

This can be called as

to(device=None, dtype=None, non_blocking=False)

to(dtype, non_blocking=False)

to(tensor, non_blocking=False)

to(memory_format=torch.channels_last)

Its signature is similar to torch.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices.

See below for examples.

Note

This method modifies the module in-place.

Parameters:

device (torch.device) – the desired device of the parameters and buffers in this module
dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module
tensor (torch.Tensor) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module
memory_format (torch.memory_format) – the desired memory format for 4D parameters and buffers in this module (keyword only argument)

Returns:

self

Return type:

Module

Examples:

>>> # xdoctest: +IGNORE_WANT("non-deterministic")
>>> linear = nn.Linear(2, 2)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]])
>>> linear.to(torch.double)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]], dtype=torch.float64)
>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_CUDA1)
>>> gpu1 = torch.device("cuda:1")
>>> linear.to(gpu1, dtype=torch.half, non_blocking=True)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16, device='cuda:1')
>>> cpu = torch.device("cpu")
>>> linear.to(cpu)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16)

>>> linear = nn.Linear(2, 2, bias=None).to(torch.cdouble)
>>> linear.weight
Parameter containing:
tensor([[ 0.3741+0.j,  0.2382+0.j],
        [ 0.5593+0.j, -0.4443+0.j]], dtype=torch.complex128)
>>> linear(torch.ones(3, 2, dtype=torch.cdouble))
tensor([[0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j]], dtype=torch.complex128)

to_empty(*, device: int | str | device | None, recurse: bool = True) → T

Move the parameters and buffers to the specified device without copying storage.

Parameters:

device (torch.device) – The desired device of the parameters and buffers in this module.
recurse (bool) – Whether parameters and buffers of submodules should be recursively moved to the specified device.

Returns:

self

Return type:

Module

train(mode: bool = True) → T

Set the module in training mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

Parameters:: mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.
Returns:: self
Return type:: Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:: dst_type (type or string) – the desired type
Returns:: self
Return type:: Module

xpu(device: int | device | None = None) → T

Move all model parameters and buffers to the XPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on XPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

zero_grad(set_to_none: bool = True) → None

Reset gradients of all model parameters.

See similar function under torch.optim.Optimizer for more context.

Parameters:: set_to_none (bool) – instead of setting to zero, set the grads to None. See torch.optim.Optimizer.zero_grad() for details.

class qsprpred.extra.gpu.models.neural_network.STFullyConnected(n_dim, n_class, device, gpus, n_epochs=100, lr=None, batch_size=256, patience=50, tol=0, is_reg=True, neurons_h1=256, neurons_hx=128, extra_layer=False, dropout_frac=0.25)[source]

Bases: Base

Single task DNN classification/regression model.

It contains four fully connected layers between which are dropout layers for robustness.

Variables:

n_dim (int) – the No. of columns (features) for input tensor
n_class (int) – the No. of columns (classes) for output tensor.
device (torch.cude) – device to run the model on
gpus (list) – list of gpu ids to run the model on
n_epochs (int) – max number of epochs
lr (float) – neural net learning rate
batch_size (int) – batch size for training
patience (int) – early stopping patience
tol (float) – early stopping tolerance
is_reg (bool) – whether the model is for regression or classification
neurons_h1 (int) – No. of neurons in the first hidden layer
neurons_hx (int) – No. of neurons in the second hidden layer
extra_layer (bool) – whether to add an extra hidden layer
dropout_frac (float) – dropout fraction
criterion (torch.nn.Module) – the loss function
dropout (torch.nn.Module) – the dropout layer
fc0 (torch.nn.Module) – the first fully connected layer
fc1 (torch.nn.Module) – the second fully connected layer
fc2 (torch.nn.Module) – the third fully connected layer
fc3 (torch.nn.Module) – the fourth fully connected layer
activation (torch.nn.Module) – the activation function

Initialize the STFullyConnected model.

Parameters:

n_dim (int) – the No. of columns (features) for input tensor
n_class (int) – the No. of columns (classes) for output tensor.
device (torch.cude) – device to run the model on
gpus (list) – list of gpu ids to run the model on
n_epochs (int) – max number of epochs
lr (float) – neural net learning rate
batch_size (int) – batch size
patience (int) – number of epochs to wait before early stop if no progress on validation set score, if patience = -1, always train to n_epochs
tol (float) – minimum absolute improvement of loss necessary to count as progress on best validation score
is_reg (bool, optional) – Regression model (True) or Classification model (False)
neurons_h1 (int) – number of neurons in first hidden layer
neurons_hx (int) – number of neurons in other hidden layers
extra_layer (bool) – add third hidden layer
dropout_frac (float) – dropout fraction

T_destination = ~T_destination

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

name (str) – name of the child module. The child module can be accessed from this module using the given name
module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:: fn (Module -> None) – function to be applied to each submodule
Returns:: self
Return type:: Module

Example:

>>> @torch.no_grad()
>>> def init_weights(m):
>>>     print(m)
>>>     if type(m) == nn.Linear:
>>>         m.weight.fill_(1.0)
>>>         print(m.weight)
>>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2))
>>> net.apply(init_weights)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:: recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.
Yields:: torch.Tensor – module buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for buf in model.buffers():
>>>     print(type(buf), buf.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

call_super_init: bool = False

children() → Iterator[Module]

Return an iterator over immediate children modules.

Yields:: Module – a child module

compile(*args, **kwargs)

Compile this Module’s forward using torch.compile().

This Module’s __call__ method is compiled and all arguments are passed as-is to torch.compile().

See torch.compile() for details on the arguments for this function.

cpu() → T

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

cuda(device: int | device | None = None) → T

Move all model parameters and buffers to the GPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on GPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

dump_patches: bool = False

eval() → T

Set the module in evaluation mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

This is equivalent with self.train(False).

See locally-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:: self
Return type:: Module

evaluate(loader) → float

Evaluate the performance of the DNN model.

Parameters:: loader (torch.util.data.DataLoader) – data loader for test set, including m X n target FloatTensor and l X n label FloatTensor (m is the No. of sample, n is the No. of features, l is the No. of classes or tasks)
Returns:: the average loss value based on the calculation of loss function with given test set.
Return type:: loss (float)

extra_repr() → str

Set the extra representation of the module.

To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.

fit(X_train, y_train, X_valid=None, y_valid=None, monitor: FitMonitor | None = None) → int

Training the DNN model.

Training is, similar to the scikit-learn or Keras style. It saves the optimal value of parameters.

Parameters:

X_train (np.ndarray or pd.Dataframe) – training data (m X n), m is the No. of samples, n is the No. of features
y_train (np.ndarray or pd.Dataframe) – training target (m X l), m is the No. of samples, l is the No. of classes or tasks
X_valid (np.ndarray or pd.Dataframe) – validation data (m X n), m is the No. of samples, n is the No. of features
y_valid (np.ndarray or pd.Dataframe) – validation target (m X l), m is the No. of samples, l is the No. of classes or tasks
monitor (FitMonitor) – monitor to use for training, if None, use base monitor

Returns:

the epoch number when the optimal model is saved

Return type:

int

float() → T

Casts all floating point parameters and buffers to float datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

forward(X, is_train=False) → Tensor[source]

Invoke the class directly as a function.

Parameters:

X (FloatTensor) – m X n FloatTensor, m is the No. of samples, n is the No. of features.
is_train (bool, optional) – is it invoked during training process (True) or just for prediction (False)

Returns:

m X n FloatTensor, m is the No. of samples,: n is the No. of classes

Return type:

y (FloatTensor)

getDataLoader(X, y=None)

Convert data to tensors and get generator over dataset with dataloader.

Parameters:

X (numpy 2d array) – input dataset
y (numpy 1d column vector) – output data

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The buffer referenced by target
Return type:: torch.Tensor
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return any extra state to include in the module’s state_dict.

Implement this and a corresponding set_extra_state() for your module if you need to store extra state. This function is called when building the module’s state_dict().

Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes.

Returns:: Any extra state to store in the module’s state_dict
Return type:: object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)
Returns:: The Parameter referenced by target
Return type:: torch.nn.Parameter
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_params(deep=True) → dict

Get parameters for this estimator.

Function copied from sklearn.base_estimator!

Parameters:: deep (bool) – If True, will return the parameters for this estimator
Returns:: Parameter names mapped to their values.
Return type:: params (dict)

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw an error.

For example, let’s say you have an nn.Module A that looks like this:

A(
    (net_b): Module(
        (net_c): Module(
            (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2))
        )
        (linear): Linear(in_features=100, out_features=200, bias=True)
    )
)

(The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used.

Parameters:: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)
Returns:: The submodule referenced by target
Return type:: torch.nn.Module
Raises:: AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

half() → T

Casts all floating point parameters and buffers to half datatype.

Note

This method modifies the module in-place.

Returns:: self
Return type:: Module

initModel()[source]: Define the layers of the model.

ipu(device: int | device | None = None) → T

Move all model parameters and buffers to the IPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on IPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

state_dict (dict) – a dict containing parameters and persistent buffers.
strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True
assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

missing_keys is a list of str containing any keys that are expected
by this module but missing from the provided state_dict.
unexpected_keys is a list of str containing the keys that are not
expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:: Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.modules()):
...     print(idx, '->', m)

0 -> Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

prefix (str) – prefix to prepend to all buffer names.
recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True.
remove_duplicate (bool, optional) – whether to remove the duplicated buffers in the result. Defaults to True.

Yields:

(str, torch.Tensor) – Tuple containing the name and buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:: (str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

memo – a memo to store the set of modules already added to the result
prefix – a prefix that will be added to the name of the module
remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, l will be returned only once.

Example:

>>> l = nn.Linear(2, 2)
>>> net = nn.Sequential(l, l)
>>> for idx, m in enumerate(net.named_modules()):
...     print(idx, '->', m)

0 -> ('', Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

prefix (str) – prefix to prepend to all parameter names.
recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:: recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.
Yields:: Parameter – module parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for param in model.parameters():
>>>     print(type(param), param.size())
<class 'torch.Tensor'> (20L,)
<class 'torch.Tensor'> (20L, 1L, 5L, 5L)

predict(X_test) → ndarray

Predicting the probability of each sample in the given dataset.

Parameters:: X_test (ndarray) – m X n target array (m is the No. of sample, n is the No. of features)
Returns:: probability of each sample in the given dataset, it is an m X l FloatTensor (m is the No. of sample, l is the No. of classes or tasks.)
Return type:: score (ndarray)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register a backward hook on the module.

This function is deprecated in favor of register_full_backward_hook() and the behavior of this function will change in future versions.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

name (str) – name of the buffer. The buffer can be accessed from this module using the given name
tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.
persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature:

hook(module, args, kwargs, output) -> None or modified output

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If True, the provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If True, the hook will be passed the kwargs given to the forward function. Default: False
always_call (bool) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_forward_pre_hook(hook: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register a forward pre-hook on the module.

The hook will be called every time before forward() is invoked.

If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature:

hook(module, args) -> None or modified input

If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature:

hook(module, args, kwargs) -> None or a tuple of modified input and kwargs

Parameters:

hook (Callable) – The user defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False
with_kwargs (bool) – If true, the hook will be passed the kwargs given to the forward function. Default: False

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward hook on the module.

The hook will be called every time the gradients with respect to a module are computed, i.e. the hook will execute if and only if the gradients with respect to module outputs are computed. The hook should have the following signature:

hook(module, grad_input, grad_output) -> tuple(Tensor) or None

The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_full_backward_pre_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register a backward pre-hook on the module.

The hook will be called every time the gradients for the module are computed. The hook should have the following signature:

hook(module, grad_output) -> tuple[Tensor] or None

The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments.

For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.

Warning

Modifying inputs inplace is not allowed when using backward hooks and will raise an error.

Parameters:

hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided hook will be fired before all existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method.

Returns:

a handle that can be used to remove the added hook by calling handle.remove()

Return type:

torch.utils.hooks.RemovableHandle

register_load_state_dict_post_hook(hook)

Register a post hook to be run after module’s load_state_dict is called.

It should have the following signature::: hook(module, incompatible_keys) -> None

The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys.

The given incompatible_keys can be modified inplace if needed.

Note that the checks performed when calling load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error.

Returns:: a handle that can be used to remove the added hook by calling handle.remove()
Return type:: torch.utils.hooks.RemovableHandle

register_module(name: str, module: Module | None) → None: Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

name (str) – name of the parameter. The parameter can be accessed from this module using the given name
param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

These hooks will be called with arguments: self, prefix, and keep_vars before calling state_dict on self. The registered hooks can be used to perform pre-processing before the state_dict call is made.

requires_grad_(requires_grad: bool = True) → T

Change if autograd should record operations on parameters in this module.

This method sets the parameters’ requires_grad attributes in-place.

This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training).

See locally-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:: requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.
Returns:: self
Return type:: Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_extra_state() for your module if you need to store extra state within its state_dict.

Parameters:: state (dict) – Extra state from the state_dict

set_params(**params) → STFullyConnected[source]

Set parameters and re-initialize model.

Parameters:: **params – parameters to be set
Returns:: the model itself
Return type:: self (STFullyConnected)

share_memory() → T: See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included.

Note

The returned object is a shallow copy. It contains references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

destination (dict, optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None.
prefix (str, optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''.
keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast the parameters and buffers.

This can be called as

to(device=None, dtype=None, non_blocking=False)

to(dtype, non_blocking=False)

to(tensor, non_blocking=False)

to(memory_format=torch.channels_last)

Its signature is similar to torch.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices.

See below for examples.

Note

This method modifies the module in-place.

Parameters:

device (torch.device) – the desired device of the parameters and buffers in this module
dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module
tensor (torch.Tensor) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module
memory_format (torch.memory_format) – the desired memory format for 4D parameters and buffers in this module (keyword only argument)

Returns:

self

Return type:

Module

Examples:

>>> # xdoctest: +IGNORE_WANT("non-deterministic")
>>> linear = nn.Linear(2, 2)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]])
>>> linear.to(torch.double)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1913, -0.3420],
        [-0.5113, -0.2325]], dtype=torch.float64)
>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_CUDA1)
>>> gpu1 = torch.device("cuda:1")
>>> linear.to(gpu1, dtype=torch.half, non_blocking=True)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16, device='cuda:1')
>>> cpu = torch.device("cpu")
>>> linear.to(cpu)
Linear(in_features=2, out_features=2, bias=True)
>>> linear.weight
Parameter containing:
tensor([[ 0.1914, -0.3420],
        [-0.5112, -0.2324]], dtype=torch.float16)

>>> linear = nn.Linear(2, 2, bias=None).to(torch.cdouble)
>>> linear.weight
Parameter containing:
tensor([[ 0.3741+0.j,  0.2382+0.j],
        [ 0.5593+0.j, -0.4443+0.j]], dtype=torch.complex128)
>>> linear(torch.ones(3, 2, dtype=torch.cdouble))
tensor([[0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j],
        [0.6122+0.j, 0.1150+0.j]], dtype=torch.complex128)

to_empty(*, device: int | str | device | None, recurse: bool = True) → T

Move the parameters and buffers to the specified device without copying storage.

Parameters:

device (torch.device) – The desired device of the parameters and buffers in this module.
recurse (bool) – Whether parameters and buffers of submodules should be recursively moved to the specified device.

Returns:

self

Return type:

Module

train(mode: bool = True) → T

Set the module in training mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

Parameters:: mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.
Returns:: self
Return type:: Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:: dst_type (type or string) – the desired type
Returns:: self
Return type:: Module

xpu(device: int | device | None = None) → T

Move all model parameters and buffers to the XPU.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on XPU while being optimized.

Note

This method modifies the module in-place.

Parameters:: device (int, optional) – if specified, all parameters will be copied to that device
Returns:: self
Return type:: Module

zero_grad(set_to_none: bool = True) → None

Reset gradients of all model parameters.

See similar function under torch.optim.Optimizer for more context.

Parameters:: set_to_none (bool) – instead of setting to zero, set the grads to None. See torch.optim.Optimizer.zero_grad() for details.

qsprpred.extra.gpu.models.tests module

class qsprpred.extra.gpu.models.tests.BenchMarkTest(methodName='runTest')[source]

Bases: BenchMarkTestCase

Test GPU models with benchmarks.

Create an instance of the class that will use the named test method when executed. Raises a ValueError if the instance does not have a method with the specified name.

classmethod addClassCleanup(function, /, *args, **kwargs): Same as addCleanup, except the cleanup items are called even if setUpClass fails (unlike tearDownClass).

addCleanup(function, /, *args, **kwargs)

Add a function, with arguments, to be called when the test is completed. Functions added are called on a LIFO basis and are called after tearDown on test failure or success.

Cleanup items are called even if setUp fails (unlike tearDown).

addTypeEqualityFunc(typeobj, function)

Add a type specific assertEqual style function to compare a type.

This method is for use by TestCase subclasses that need to register their own type equality functions to provide nicer error messages.

Parameters:

typeobj – The data type to call this function on when both values are of the same type in assertEqual().
function – The callable taking two arguments and an optional msg= argument that raises self.failureException with a useful error message when the two arguments are not equal.

assertAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are unequal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is more than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

If the two objects compare equal then they will automatically compare almost equal.

assertCountEqual(first, second, msg=None)

Asserts that two iterables have the same elements, the same number of times, without regard to order.

self.assertEqual(Counter(list(first)),
Counter(list(second)))

Example:

[0, 1, 1] and [1, 0, 1] compare equal.

[0, 0, 1] and [0, 1] compare unequal.

assertDictEqual(d1, d2, msg=None)

assertEqual(first, second, msg=None): Fail if the two objects are unequal as determined by the ‘==’ operator.

assertFalse(expr, msg=None): Check that the expression is false.

assertGreater(a, b, msg=None): Just like self.assertTrue(a > b), but with a nicer default message.

assertGreaterEqual(a, b, msg=None): Just like self.assertTrue(a >= b), but with a nicer default message.

assertIn(member, container, msg=None): Just like self.assertTrue(a in b), but with a nicer default message.

assertIs(expr1, expr2, msg=None): Just like self.assertTrue(a is b), but with a nicer default message.

assertIsInstance(obj, cls, msg=None): Same as self.assertTrue(isinstance(obj, cls)), with a nicer default message.

assertIsNone(obj, msg=None): Same as self.assertTrue(obj is None), with a nicer default message.

assertIsNot(expr1, expr2, msg=None): Just like self.assertTrue(a is not b), but with a nicer default message.

assertIsNotNone(obj, msg=None): Included for symmetry with assertIsNone.

assertLess(a, b, msg=None): Just like self.assertTrue(a < b), but with a nicer default message.

assertLessEqual(a, b, msg=None): Just like self.assertTrue(a <= b), but with a nicer default message.

assertListEqual(list1, list2, msg=None)

A list-specific equality assertion.

Parameters:

list1 – The first list to compare.
list2 – The second list to compare.
msg – Optional message to use on failure instead of a list of differences.

assertLogs(logger=None, level=None)

Fail unless a log message of level level or higher is emitted on logger_name or its children. If omitted, level defaults to INFO and logger defaults to the root logger.

This method must be used as a context manager, and will yield a recording object with two attributes: output and records. At the end of the context manager, the output attribute will be a list of the matching formatted log messages and the records attribute will be a list of the corresponding LogRecord objects.

Example:

with self.assertLogs('foo', level='INFO') as cm:
    logging.getLogger('foo').info('first message')
    logging.getLogger('foo.bar').error('second message')
self.assertEqual(cm.output, ['INFO:foo:first message',
                             'ERROR:foo.bar:second message'])

assertMultiLineEqual(first, second, msg=None): Assert that two multi-line strings are equal.

assertNoLogs(logger=None, level=None)

Fail unless no log messages of level level or higher are emitted on logger_name or its children.

This method must be used as a context manager.

assertNotAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are equal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is less than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

Objects that are equal automatically fail.

assertNotEqual(first, second, msg=None): Fail if the two objects are equal as determined by the ‘!=’ operator.

assertNotIn(member, container, msg=None): Just like self.assertTrue(a not in b), but with a nicer default message.

assertNotIsInstance(obj, cls, msg=None): Included for symmetry with assertIsInstance.

assertNotRegex(text, unexpected_regex, msg=None): Fail the test if the text matches the regular expression.

assertRaises(expected_exception, *args, **kwargs)

Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments. If a different type of exception is raised, it will not be caught, and the test case will be deemed to have suffered an error, exactly as for an unexpected exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertRaises(SomeException):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertRaises is used as a context object.

The context manager keeps a reference to the exception as the ‘exception’ attribute. This allows you to inspect the exception after the assertion:

with self.assertRaises(SomeException) as cm:
    do_something()
the_exception = cm.exception
self.assertEqual(the_exception.error_code, 3)

assertRaisesRegex(expected_exception, expected_regex, *args, **kwargs)

Asserts that the message in a raised exception matches a regex.

Parameters:

expected_exception – Exception class expected to be raised.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertRaisesRegex is used as a context manager.

assertRegex(text, expected_regex, msg=None): Fail the test unless the text matches the regular expression.

assertSequenceEqual(seq1, seq2, msg=None, seq_type=None)

An equality assertion for ordered sequences (like lists and tuples).

For the purposes of this function, a valid ordered sequence type is one which can be indexed, has a length, and has an equality operator.

Parameters:

seq1 – The first sequence to compare.
seq2 – The second sequence to compare.
seq_type – The expected datatype of the sequences, or None if no datatype should be enforced.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual(set1, set2, msg=None)

A set-specific equality assertion.

Parameters:

set1 – The first set to compare.
set2 – The second set to compare.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual uses ducktyping to support different types of sets, and is optimized for sets specifically (parameters must support a difference method).

assertTrue(expr, msg=None): Check that the expression is true.

assertTupleEqual(tuple1, tuple2, msg=None)

A tuple-specific equality assertion.

Parameters:

tuple1 – The first tuple to compare.
tuple2 – The second tuple to compare.
msg – Optional message to use on failure instead of a list of differences.

assertWarns(expected_warning, *args, **kwargs)

Fail unless a warning of class warnClass is triggered by the callable when invoked with specified positional and keyword arguments. If a different type of warning is triggered, it will not be handled: depending on the other warning filtering rules in effect, it might be silenced, printed out, or raised as an exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertWarns(SomeWarning):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertWarns is used as a context object.

The context manager keeps a reference to the first matching warning as the ‘warning’ attribute; similarly, the ‘filename’ and ‘lineno’ attributes give you information about the line of Python code from which the warning was triggered. This allows you to inspect the warning after the assertion:

with self.assertWarns(SomeWarning) as cm:
    do_something()
the_warning = cm.warning
self.assertEqual(the_warning.some_attribute, 147)

assertWarnsRegex(expected_warning, expected_regex, *args, **kwargs)

Asserts that the message in a triggered warning matches a regexp. Basic functioning is similar to assertWarns() with the addition that only warnings whose messages also match the regular expression are considered successful matches.

Parameters:

expected_warning – Warning class expected to be triggered.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertWarnsRegex is used as a context manager.

checkRunResults(results)

checkSettings()

clearGenerated(): Remove the directories that are used for testing.

countTestCases()

createLargeMultitaskDataSet(name='QSPRDataset_multi_test', target_props=[{'name': 'HBD', 'task': <TargetTasks.MULTICLASS: 'MULTICLASS'>, 'th': [-1, 1, 2, 100]}, {'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
preparation_settings (dict) – dictionary containing preparation settings
random_state (int) – random state to use for splitting and shuffling

Returns:

a QSPRDataset object

Return type:

createLargeTestDataSet(name='QSPRDataset_test_large', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42, n_jobs=1, chunk_size=None)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createSmallTestDataSet(name='QSPRDataset_test_small', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a small dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createTestDataSetFromFrame(df, name='QSPRDataset_test', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], random_state=None, prep=None, n_jobs=1, chunk_size=None)

Create a dataset for testing purposes from the given data frame.

Parameters:

df (pd.DataFrame) – data frame containing the dataset
name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
prep (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

debug(): Run the test without collecting errors in a TestResult

defaultTestResult()

classmethod doClassCleanups(): Execute all class cleanup functions. Normally called for you after tearDownClass.

doCleanups(): Execute all cleanup functions. Normally called for you after tearDown.

classmethod enterClassContext(cm): Same as enterContext, but class-wide.

enterContext(cm)

Enters the supplied context manager.

If successful, also adds its __exit__ method as a cleanup function and returns the result of the __enter__ method.

fail(msg=None): Fail immediately, with the given message.

failureException: alias of AssertionError

classmethod getAllDescriptors()

Return a list of (ideally) all available descriptor sets. For now they need to be added manually to the list below.

TODO: would be nice to create the list automatically by implementing a descriptor set registry that would hold all installed descriptor sets.

Returns:: list of DescriptorCalculator objects
Return type:: list

getBigDF()

Get a large data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

classmethod getDataPrepGrid()

Return a list of many possible combinations of descriptor calculators, splits, feature standardizers, feature filters and data filters. Again, this is not exhaustive, but should cover a lot of cases.

Returns:: a generator that yields tuples of all possible combinations as stated above, each tuple is defined as: (descriptor_calculator, split, feature_standardizer, feature_filters, data_filters)
Return type:: grid

classmethod getDefaultCalculatorCombo()

Makes a list of default descriptor calculators that can be used in tests. It creates a calculator with only morgan fingerprints and rdkit descriptors, but also one with them both to test behaviour with multiple descriptor sets. Override this method if you want to test with other descriptor sets and calculator combinations.

Returns:: list of created DescriptorCalculator objects
Return type:: list

static getDefaultPrep(): Return a dictionary with default preparation settings.

classmethod getPrepCombos()

Return a list of all possible preparation combinations as generated by getDataPrepGrid as well as their names. The generated list can be used to parameterize tests with the given named combinations.

Returns:: list of `list`s of all possible combinations of preparation
Return type:: list

getSmallDF()

Get a small data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

id()

longMessage = True

maxDiff = 640

run(result=None)

setUp(): Hook method for setting up the test fixture before exercising it.

classmethod setUpClass(): Hook method for setting up class fixture before running tests in the class.

setUpPaths(): Create the directories that are used for testing.

shortDescription()

Returns a one-line description of the test, or None if no description has been provided.

The default implementation of this method returns the first line of the specified test method’s docstring.

skipTest(reason): Skip this test.

subTest(msg=<object object>, **params): Return a context manager that will return the enclosed block of code in a subtest identified by the optional message and keyword parameters. A failure in the subtest marks the test case as failed but resumes execution at the end of the enclosed block, allowing further test code to be executed.

tearDown(): Remove all files and directories that are used for testing.

classmethod tearDownClass(): Hook method for deconstructing the class fixture after running all tests in the class.

testBasicTorchExecution()[source]: Run single task tests for classification.

testChemProp()[source]: Run single task tests for classification.

validate_split(dataset): Check if the split has the data it should have after splitting.

class qsprpred.extra.gpu.models.tests.ChemPropTest(methodName='runTest')[source]

Bases: ModelDataSetsPathMixIn, ModelCheckMixIn, TestCase

This class holds the tests for the DNNModel class.

Create an instance of the class that will use the named test method when executed. Raises a ValueError if the instance does not have a method with the specified name.

classmethod addClassCleanup(function, /, *args, **kwargs): Same as addCleanup, except the cleanup items are called even if setUpClass fails (unlike tearDownClass).

addCleanup(function, /, *args, **kwargs)

Add a function, with arguments, to be called when the test is completed. Functions added are called on a LIFO basis and are called after tearDown on test failure or success.

Cleanup items are called even if setUp fails (unlike tearDown).

addTypeEqualityFunc(typeobj, function)

Add a type specific assertEqual style function to compare a type.

This method is for use by TestCase subclasses that need to register their own type equality functions to provide nicer error messages.

Parameters:

typeobj – The data type to call this function on when both values are of the same type in assertEqual().
function – The callable taking two arguments and an optional msg= argument that raises self.failureException with a useful error message when the two arguments are not equal.

assertAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are unequal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is more than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

If the two objects compare equal then they will automatically compare almost equal.

assertCountEqual(first, second, msg=None)

Asserts that two iterables have the same elements, the same number of times, without regard to order.

self.assertEqual(Counter(list(first)),
Counter(list(second)))

Example:

[0, 1, 1] and [1, 0, 1] compare equal.

[0, 0, 1] and [0, 1] compare unequal.

assertDictEqual(d1, d2, msg=None)

assertEqual(first, second, msg=None): Fail if the two objects are unequal as determined by the ‘==’ operator.

assertFalse(expr, msg=None): Check that the expression is false.

assertGreater(a, b, msg=None): Just like self.assertTrue(a > b), but with a nicer default message.

assertGreaterEqual(a, b, msg=None): Just like self.assertTrue(a >= b), but with a nicer default message.

assertIn(member, container, msg=None): Just like self.assertTrue(a in b), but with a nicer default message.

assertIs(expr1, expr2, msg=None): Just like self.assertTrue(a is b), but with a nicer default message.

assertIsInstance(obj, cls, msg=None): Same as self.assertTrue(isinstance(obj, cls)), with a nicer default message.

assertIsNone(obj, msg=None): Same as self.assertTrue(obj is None), with a nicer default message.

assertIsNot(expr1, expr2, msg=None): Just like self.assertTrue(a is not b), but with a nicer default message.

assertIsNotNone(obj, msg=None): Included for symmetry with assertIsNone.

assertLess(a, b, msg=None): Just like self.assertTrue(a < b), but with a nicer default message.

assertLessEqual(a, b, msg=None): Just like self.assertTrue(a <= b), but with a nicer default message.

assertListEqual(list1, list2, msg=None)

A list-specific equality assertion.

Parameters:

list1 – The first list to compare.
list2 – The second list to compare.
msg – Optional message to use on failure instead of a list of differences.

assertLogs(logger=None, level=None)

Fail unless a log message of level level or higher is emitted on logger_name or its children. If omitted, level defaults to INFO and logger defaults to the root logger.

This method must be used as a context manager, and will yield a recording object with two attributes: output and records. At the end of the context manager, the output attribute will be a list of the matching formatted log messages and the records attribute will be a list of the corresponding LogRecord objects.

Example:

with self.assertLogs('foo', level='INFO') as cm:
    logging.getLogger('foo').info('first message')
    logging.getLogger('foo.bar').error('second message')
self.assertEqual(cm.output, ['INFO:foo:first message',
                             'ERROR:foo.bar:second message'])

assertMultiLineEqual(first, second, msg=None): Assert that two multi-line strings are equal.

assertNoLogs(logger=None, level=None)

Fail unless no log messages of level level or higher are emitted on logger_name or its children.

This method must be used as a context manager.

assertNotAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are equal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is less than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

Objects that are equal automatically fail.

assertNotEqual(first, second, msg=None): Fail if the two objects are equal as determined by the ‘!=’ operator.

assertNotIn(member, container, msg=None): Just like self.assertTrue(a not in b), but with a nicer default message.

assertNotIsInstance(obj, cls, msg=None): Included for symmetry with assertIsInstance.

assertNotRegex(text, unexpected_regex, msg=None): Fail the test if the text matches the regular expression.

assertRaises(expected_exception, *args, **kwargs)

Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments. If a different type of exception is raised, it will not be caught, and the test case will be deemed to have suffered an error, exactly as for an unexpected exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertRaises(SomeException):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertRaises is used as a context object.

The context manager keeps a reference to the exception as the ‘exception’ attribute. This allows you to inspect the exception after the assertion:

with self.assertRaises(SomeException) as cm:
    do_something()
the_exception = cm.exception
self.assertEqual(the_exception.error_code, 3)

assertRaisesRegex(expected_exception, expected_regex, *args, **kwargs)

Asserts that the message in a raised exception matches a regex.

Parameters:

expected_exception – Exception class expected to be raised.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertRaisesRegex is used as a context manager.

assertRegex(text, expected_regex, msg=None): Fail the test unless the text matches the regular expression.

assertSequenceEqual(seq1, seq2, msg=None, seq_type=None)

An equality assertion for ordered sequences (like lists and tuples).

For the purposes of this function, a valid ordered sequence type is one which can be indexed, has a length, and has an equality operator.

Parameters:

seq1 – The first sequence to compare.
seq2 – The second sequence to compare.
seq_type – The expected datatype of the sequences, or None if no datatype should be enforced.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual(set1, set2, msg=None)

A set-specific equality assertion.

Parameters:

set1 – The first set to compare.
set2 – The second set to compare.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual uses ducktyping to support different types of sets, and is optimized for sets specifically (parameters must support a difference method).

assertTrue(expr, msg=None): Check that the expression is true.

assertTupleEqual(tuple1, tuple2, msg=None)

A tuple-specific equality assertion.

Parameters:

tuple1 – The first tuple to compare.
tuple2 – The second tuple to compare.
msg – Optional message to use on failure instead of a list of differences.

assertWarns(expected_warning, *args, **kwargs)

Fail unless a warning of class warnClass is triggered by the callable when invoked with specified positional and keyword arguments. If a different type of warning is triggered, it will not be handled: depending on the other warning filtering rules in effect, it might be silenced, printed out, or raised as an exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertWarns(SomeWarning):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertWarns is used as a context object.

The context manager keeps a reference to the first matching warning as the ‘warning’ attribute; similarly, the ‘filename’ and ‘lineno’ attributes give you information about the line of Python code from which the warning was triggered. This allows you to inspect the warning after the assertion:

with self.assertWarns(SomeWarning) as cm:
    do_something()
the_warning = cm.warning
self.assertEqual(the_warning.some_attribute, 147)

assertWarnsRegex(expected_warning, expected_regex, *args, **kwargs)

Asserts that the message in a triggered warning matches a regexp. Basic functioning is similar to assertWarns() with the addition that only warnings whose messages also match the regular expression are considered successful matches.

Parameters:

expected_warning – Warning class expected to be triggered.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertWarnsRegex is used as a context manager.

checkOptimization(model: QSPRModel, ds: QSPRDataset, optimizer: HyperparameterOptimization)

clearGenerated(): Remove the directories that are used for testing.

countTestCases()

createLargeMultitaskDataSet(name='QSPRDataset_multi_test', target_props=[{'name': 'HBD', 'task': <TargetTasks.MULTICLASS: 'MULTICLASS'>, 'th': [-1, 1, 2, 100]}, {'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
preparation_settings (dict) – dictionary containing preparation settings
random_state (int) – random state to use for splitting and shuffling

Returns:

a QSPRDataset object

Return type:

createLargeTestDataSet(name='QSPRDataset_test_large', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42, n_jobs=1, chunk_size=None)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createSmallTestDataSet(name='QSPRDataset_test_small', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a small dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createTestDataSetFromFrame(df, name='QSPRDataset_test', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], random_state=None, prep=None, n_jobs=1, chunk_size=None)

Create a dataset for testing purposes from the given data frame.

Parameters:

df (pd.DataFrame) – data frame containing the dataset
name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
prep (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

debug(): Run the test without collecting errors in a TestResult

defaultTestResult()

classmethod doClassCleanups(): Execute all class cleanup functions. Normally called for you after tearDownClass.

doCleanups(): Execute all cleanup functions. Normally called for you after tearDown.

classmethod enterClassContext(cm): Same as enterContext, but class-wide.

enterContext(cm)

Enters the supplied context manager.

If successful, also adds its __exit__ method as a cleanup function and returns the result of the __enter__ method.

fail(msg=None): Fail immediately, with the given message.

failureException: alias of AssertionError

fitTest(model: QSPRModel, ds: QSPRDataset)

Test model fitting, optimization and evaluation.

Parameters:

model (QSPRModel) – The model to test.
ds (QSPRDataset) – The dataset to use for testing.

classmethod getAllDescriptors()

Return a list of (ideally) all available descriptor sets. For now they need to be added manually to the list below.

TODO: would be nice to create the list automatically by implementing a descriptor set registry that would hold all installed descriptor sets.

Returns:: list of DescriptorCalculator objects
Return type:: list

getBigDF()

Get a large data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

classmethod getDataPrepGrid()

Return a list of many possible combinations of descriptor calculators, splits, feature standardizers, feature filters and data filters. Again, this is not exhaustive, but should cover a lot of cases.

Returns:: a generator that yields tuples of all possible combinations as stated above, each tuple is defined as: (descriptor_calculator, split, feature_standardizer, feature_filters, data_filters)
Return type:: grid

classmethod getDefaultCalculatorCombo()

Makes a list of default descriptor calculators that can be used in tests. It creates a calculator with only morgan fingerprints and rdkit descriptors, but also one with them both to test behaviour with multiple descriptor sets. Override this method if you want to test with other descriptor sets and calculator combinations.

Returns:: list of created DescriptorCalculator objects
Return type:: list

static getDefaultPrep(): Return a dictionary with default preparation settings.

getModel(name: str, parameters: dict | None = None, random_state: int | None = None)[source]

Initialize model with data set.

Parameters:

name – Name of the model.
parameters – Parameters to use.
random_state – Random seed to use for random operations.

getParamGrid(model: QSPRModel, grid: str) → dict

Get the parameter grid for a model.

Parameters:

model (QSPRModel) – The model to get the parameter grid for.
grid (str) – The grid type to get the parameter grid for.

Returns:

The parameter grid.

Return type:

classmethod getPrepCombos()

Return a list of all possible preparation combinations as generated by getDataPrepGrid as well as their names. The generated list can be used to parameterize tests with the given named combinations.

Returns:: list of `list`s of all possible combinations of preparation
Return type:: list

getSmallDF()

Get a small data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

property gridFile: Return the path to the grid file with test search spaces for hyperparameter optimization.

id()

longMessage = True

maxDiff = 640

predictorTest(model: QSPRModel, dataset: QSPRDataset, comparison_model: QSPRModel | None = None, expect_equal_result=True, **pred_kwargs)

Test model predictions.

Checks if the shape of the predictions is as expected and if the predictions of the predictMols function are consistent with the predictions of the predict/predictProba functions. Also checks if the predictions of the model are the same as the predictions of the comparison model if given.

Parameters:

model (QSPRModel) – The model to make predictions with.
dataset (QSPRDataset) – The dataset to make predictions for.
comparison_model (QSPRModel) – another model to compare the predictions with.
expect_equal_result (bool) – Whether the expected result should be equal or not equal to the predictions of the comparison model.
**pred_kwargs – Extra keyword arguments to pass to the predictor’s predictMols method.

run(result=None)

setUp()[source]: Hook method for setting up the test fixture before exercising it.

classmethod setUpClass(): Hook method for setting up class fixture before running tests in the class.

setUpPaths(): Set up the test environment.

shortDescription()

Returns a one-line description of the test, or None if no description has been provided.

The default implementation of this method returns the first line of the specified test method’s docstring.

skipTest(reason): Skip this test.

subTest(msg=<object object>, **params): Return a context manager that will return the enclosed block of code in a subtest identified by the optional message and keyword parameters. A failure in the subtest marks the test case as failed but resumes execution at the end of the enclosed block, allowing further test code to be executed.

tearDown(): Remove all files and directories that are used for testing.

classmethod tearDownClass(): Hook method for deconstructing the class fixture after running all tests in the class.

testConsistency()[source]: Test if QSPRpred Chemprop and Chemprop models are consistent.

testMultiTaskmodel = None

testMultiTaskmodel_0_MoleculeModel_MULTITASK_REGRESSION(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_MULTITASK_REGRESSION’, task=<ModelTasks.MULTITASK_REGRESSION: ‘MULTITASK_REGRESSION’>, alg_name=’MoleculeModel’, random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.

testMultiTaskmodel_1_MoleculeModel_MULTITASK_SINGLECLASS_None(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_MULTITASK_SINGLECLASS_None’, task=<ModelTasks.MULTITASK_SINGLECLASS: ‘MULTITASK_SINGLECLASS’>, alg_name=’MoleculeModel’, random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.

testMultiTaskmodel_2_MoleculeModel_MULTITASK_SINGLECLASS_1_42(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_MULTITASK_SINGLECLASS_1_42’, task=<ModelTasks.MULTITASK_SINGLECLASS: ‘MULTITASK_SINGLECLASS’>, alg_name=’MoleculeModel’, random_state=[1, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.

testMultiTaskmodel_3_MoleculeModel_MULTITASK_SINGLECLASS_42_42(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_MULTITASK_SINGLECLASS_42_42’, task=<ModelTasks.MULTITASK_SINGLECLASS: ‘MULTITASK_SINGLECLASS’>, alg_name=’MoleculeModel’, random_state=[42, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.

testSingleTaskModel = None

testSingleTaskModel_0_MoleculeModel_SINGLECLASS(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_SINGLECLASS’, task=<TargetTasks.SINGLECLASS: ‘SINGLECLASS’>, alg_name=’MoleculeModel’, th=[6.5], random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_1_MoleculeModel_MULTICLASS(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_MULTICLASS’, task=<TargetTasks.MULTICLASS: ‘MULTICLASS’>, alg_name=’MoleculeModel’, th=[0, 1, 10, 1100], random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_2_MoleculeModel_REGRESSION_None(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_REGRESSION_None’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’MoleculeModel’, th=None, random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_3_MoleculeModel_REGRESSION_1_42(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_REGRESSION_1_42’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’MoleculeModel’, th=None, random_state=[1, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_4_MoleculeModel_REGRESSION_42_42(**kw)

Test the DNNModel model in one configuration [with _=’MoleculeModel_REGRESSION_42_42’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’MoleculeModel’, th=None, random_state=[42, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

validate_split(dataset): Check if the split has the data it should have after splitting.

class qsprpred.extra.gpu.models.tests.NeuralNet(methodName='runTest')[source]

Bases: ModelDataSetsPathMixIn, ModelCheckMixIn, TestCase

This class holds the tests for the DNNModel class.

Create an instance of the class that will use the named test method when executed. Raises a ValueError if the instance does not have a method with the specified name.

classmethod addClassCleanup(function, /, *args, **kwargs): Same as addCleanup, except the cleanup items are called even if setUpClass fails (unlike tearDownClass).

addCleanup(function, /, *args, **kwargs)

Add a function, with arguments, to be called when the test is completed. Functions added are called on a LIFO basis and are called after tearDown on test failure or success.

Cleanup items are called even if setUp fails (unlike tearDown).

addTypeEqualityFunc(typeobj, function)

Add a type specific assertEqual style function to compare a type.

This method is for use by TestCase subclasses that need to register their own type equality functions to provide nicer error messages.

Parameters:

typeobj – The data type to call this function on when both values are of the same type in assertEqual().
function – The callable taking two arguments and an optional msg= argument that raises self.failureException with a useful error message when the two arguments are not equal.

assertAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are unequal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is more than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

If the two objects compare equal then they will automatically compare almost equal.

assertCountEqual(first, second, msg=None)

Asserts that two iterables have the same elements, the same number of times, without regard to order.

self.assertEqual(Counter(list(first)),
Counter(list(second)))

Example:

[0, 1, 1] and [1, 0, 1] compare equal.

[0, 0, 1] and [0, 1] compare unequal.

assertDictEqual(d1, d2, msg=None)

assertEqual(first, second, msg=None): Fail if the two objects are unequal as determined by the ‘==’ operator.

assertFalse(expr, msg=None): Check that the expression is false.

assertGreater(a, b, msg=None): Just like self.assertTrue(a > b), but with a nicer default message.

assertGreaterEqual(a, b, msg=None): Just like self.assertTrue(a >= b), but with a nicer default message.

assertIn(member, container, msg=None): Just like self.assertTrue(a in b), but with a nicer default message.

assertIs(expr1, expr2, msg=None): Just like self.assertTrue(a is b), but with a nicer default message.

assertIsInstance(obj, cls, msg=None): Same as self.assertTrue(isinstance(obj, cls)), with a nicer default message.

assertIsNone(obj, msg=None): Same as self.assertTrue(obj is None), with a nicer default message.

assertIsNot(expr1, expr2, msg=None): Just like self.assertTrue(a is not b), but with a nicer default message.

assertIsNotNone(obj, msg=None): Included for symmetry with assertIsNone.

assertLess(a, b, msg=None): Just like self.assertTrue(a < b), but with a nicer default message.

assertLessEqual(a, b, msg=None): Just like self.assertTrue(a <= b), but with a nicer default message.

assertListEqual(list1, list2, msg=None)

A list-specific equality assertion.

Parameters:

list1 – The first list to compare.
list2 – The second list to compare.
msg – Optional message to use on failure instead of a list of differences.

assertLogs(logger=None, level=None)

Fail unless a log message of level level or higher is emitted on logger_name or its children. If omitted, level defaults to INFO and logger defaults to the root logger.

This method must be used as a context manager, and will yield a recording object with two attributes: output and records. At the end of the context manager, the output attribute will be a list of the matching formatted log messages and the records attribute will be a list of the corresponding LogRecord objects.

Example:

with self.assertLogs('foo', level='INFO') as cm:
    logging.getLogger('foo').info('first message')
    logging.getLogger('foo.bar').error('second message')
self.assertEqual(cm.output, ['INFO:foo:first message',
                             'ERROR:foo.bar:second message'])

assertMultiLineEqual(first, second, msg=None): Assert that two multi-line strings are equal.

assertNoLogs(logger=None, level=None)

Fail unless no log messages of level level or higher are emitted on logger_name or its children.

This method must be used as a context manager.

assertNotAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are equal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is less than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

Objects that are equal automatically fail.

assertNotEqual(first, second, msg=None): Fail if the two objects are equal as determined by the ‘!=’ operator.

assertNotIn(member, container, msg=None): Just like self.assertTrue(a not in b), but with a nicer default message.

assertNotIsInstance(obj, cls, msg=None): Included for symmetry with assertIsInstance.

assertNotRegex(text, unexpected_regex, msg=None): Fail the test if the text matches the regular expression.

assertRaises(expected_exception, *args, **kwargs)

Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments. If a different type of exception is raised, it will not be caught, and the test case will be deemed to have suffered an error, exactly as for an unexpected exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertRaises(SomeException):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertRaises is used as a context object.

The context manager keeps a reference to the exception as the ‘exception’ attribute. This allows you to inspect the exception after the assertion:

with self.assertRaises(SomeException) as cm:
    do_something()
the_exception = cm.exception
self.assertEqual(the_exception.error_code, 3)

assertRaisesRegex(expected_exception, expected_regex, *args, **kwargs)

Asserts that the message in a raised exception matches a regex.

Parameters:

expected_exception – Exception class expected to be raised.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertRaisesRegex is used as a context manager.

assertRegex(text, expected_regex, msg=None): Fail the test unless the text matches the regular expression.

assertSequenceEqual(seq1, seq2, msg=None, seq_type=None)

An equality assertion for ordered sequences (like lists and tuples).

For the purposes of this function, a valid ordered sequence type is one which can be indexed, has a length, and has an equality operator.

Parameters:

seq1 – The first sequence to compare.
seq2 – The second sequence to compare.
seq_type – The expected datatype of the sequences, or None if no datatype should be enforced.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual(set1, set2, msg=None)

A set-specific equality assertion.

Parameters:

set1 – The first set to compare.
set2 – The second set to compare.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual uses ducktyping to support different types of sets, and is optimized for sets specifically (parameters must support a difference method).

assertTrue(expr, msg=None): Check that the expression is true.

assertTupleEqual(tuple1, tuple2, msg=None)

A tuple-specific equality assertion.

Parameters:

tuple1 – The first tuple to compare.
tuple2 – The second tuple to compare.
msg – Optional message to use on failure instead of a list of differences.

assertWarns(expected_warning, *args, **kwargs)

Fail unless a warning of class warnClass is triggered by the callable when invoked with specified positional and keyword arguments. If a different type of warning is triggered, it will not be handled: depending on the other warning filtering rules in effect, it might be silenced, printed out, or raised as an exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertWarns(SomeWarning):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertWarns is used as a context object.

The context manager keeps a reference to the first matching warning as the ‘warning’ attribute; similarly, the ‘filename’ and ‘lineno’ attributes give you information about the line of Python code from which the warning was triggered. This allows you to inspect the warning after the assertion:

with self.assertWarns(SomeWarning) as cm:
    do_something()
the_warning = cm.warning
self.assertEqual(the_warning.some_attribute, 147)

assertWarnsRegex(expected_warning, expected_regex, *args, **kwargs)

Asserts that the message in a triggered warning matches a regexp. Basic functioning is similar to assertWarns() with the addition that only warnings whose messages also match the regular expression are considered successful matches.

Parameters:

expected_warning – Warning class expected to be triggered.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertWarnsRegex is used as a context manager.

checkOptimization(model: QSPRModel, ds: QSPRDataset, optimizer: HyperparameterOptimization)

clearGenerated(): Remove the directories that are used for testing.

countTestCases()

createLargeMultitaskDataSet(name='QSPRDataset_multi_test', target_props=[{'name': 'HBD', 'task': <TargetTasks.MULTICLASS: 'MULTICLASS'>, 'th': [-1, 1, 2, 100]}, {'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
preparation_settings (dict) – dictionary containing preparation settings
random_state (int) – random state to use for splitting and shuffling

Returns:

a QSPRDataset object

Return type:

createLargeTestDataSet(name='QSPRDataset_test_large', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42, n_jobs=1, chunk_size=None)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createSmallTestDataSet(name='QSPRDataset_test_small', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a small dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createTestDataSetFromFrame(df, name='QSPRDataset_test', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], random_state=None, prep=None, n_jobs=1, chunk_size=None)

Create a dataset for testing purposes from the given data frame.

Parameters:

df (pd.DataFrame) – data frame containing the dataset
name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
prep (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

debug(): Run the test without collecting errors in a TestResult

defaultTestResult()

classmethod doClassCleanups(): Execute all class cleanup functions. Normally called for you after tearDownClass.

doCleanups(): Execute all cleanup functions. Normally called for you after tearDown.

classmethod enterClassContext(cm): Same as enterContext, but class-wide.

enterContext(cm)

Enters the supplied context manager.

If successful, also adds its __exit__ method as a cleanup function and returns the result of the __enter__ method.

fail(msg=None): Fail immediately, with the given message.

failureException: alias of AssertionError

fitTest(model: QSPRModel, ds: QSPRDataset)

Test model fitting, optimization and evaluation.

Parameters:

model (QSPRModel) – The model to test.
ds (QSPRDataset) – The dataset to use for testing.

classmethod getAllDescriptors()

Return a list of (ideally) all available descriptor sets. For now they need to be added manually to the list below.

TODO: would be nice to create the list automatically by implementing a descriptor set registry that would hold all installed descriptor sets.

Returns:: list of DescriptorCalculator objects
Return type:: list

getBigDF()

Get a large data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

classmethod getDataPrepGrid()

Return a list of many possible combinations of descriptor calculators, splits, feature standardizers, feature filters and data filters. Again, this is not exhaustive, but should cover a lot of cases.

Returns:: a generator that yields tuples of all possible combinations as stated above, each tuple is defined as: (descriptor_calculator, split, feature_standardizer, feature_filters, data_filters)
Return type:: grid

classmethod getDefaultCalculatorCombo()

Makes a list of default descriptor calculators that can be used in tests. It creates a calculator with only morgan fingerprints and rdkit descriptors, but also one with them both to test behaviour with multiple descriptor sets. Override this method if you want to test with other descriptor sets and calculator combinations.

Returns:: list of created DescriptorCalculator objects
Return type:: list

static getDefaultPrep(): Return a dictionary with default preparation settings.

getModel(name: str, alg: Type | None = None, parameters: dict | None = None, random_state: int | None = None)[source]

Initialize model with data set.

Parameters:

name – Name of the model.
alg – Algorithm to use.
dataset – Data set to use.
parameters – Parameters to use.
random_state – Random seed to use for random operations.

getParamGrid(model: QSPRModel, grid: str) → dict

Get the parameter grid for a model.

Parameters:

model (QSPRModel) – The model to get the parameter grid for.
grid (str) – The grid type to get the parameter grid for.

Returns:

The parameter grid.

Return type:

classmethod getPrepCombos()

Return a list of all possible preparation combinations as generated by getDataPrepGrid as well as their names. The generated list can be used to parameterize tests with the given named combinations.

Returns:: list of `list`s of all possible combinations of preparation
Return type:: list

getSmallDF()

Get a small data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

property gridFile: Return the path to the grid file with test search spaces for hyperparameter optimization.

id()

longMessage = True

maxDiff = 640

predictorTest(model: QSPRModel, dataset: QSPRDataset, comparison_model: QSPRModel | None = None, expect_equal_result=True, **pred_kwargs)

Test model predictions.

Checks if the shape of the predictions is as expected and if the predictions of the predictMols function are consistent with the predictions of the predict/predictProba functions. Also checks if the predictions of the model are the same as the predictions of the comparison model if given.

Parameters:

model (QSPRModel) – The model to make predictions with.
dataset (QSPRDataset) – The dataset to make predictions for.
comparison_model (QSPRModel) – another model to compare the predictions with.
expect_equal_result (bool) – Whether the expected result should be equal or not equal to the predictions of the comparison model.
**pred_kwargs – Extra keyword arguments to pass to the predictor’s predictMols method.

run(result=None)

setUp()[source]: Set up the test case.

classmethod setUpClass(): Hook method for setting up class fixture before running tests in the class.

setUpPaths(): Set up the test environment.

shortDescription()

Returns a one-line description of the test, or None if no description has been provided.

The default implementation of this method returns the first line of the specified test method’s docstring.

skipTest(reason): Skip this test.

subTest(msg=<object object>, **params): Return a context manager that will return the enclosed block of code in a subtest identified by the optional message and keyword parameters. A failure in the subtest marks the test case as failed but resumes execution at the end of the enclosed block, allowing further test code to be executed.

tearDown(): Remove all files and directories that are used for testing.

classmethod tearDownClass(): Hook method for deconstructing the class fixture after running all tests in the class.

testSingleTaskModel = None

testSingleTaskModel_0_STFullyConnected_SINGLECLASS(**kw)

Test the DNNModel model in one configuration [with _=’STFullyConnected_SINGLECLASS’, task=<TargetTasks.SINGLECLASS: ‘SINGLECLASS’>, alg_name=’STFullyConnected’, alg=<class ‘qsprpred.extra.gpu.model…eural_network.STFullyConnected’>, th=[6.5], random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
alg – Algorithm to use.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_1_STFullyConnected_MULTICLASS(**kw)

Test the DNNModel model in one configuration [with _=’STFullyConnected_MULTICLASS’, task=<TargetTasks.MULTICLASS: ‘MULTICLASS’>, alg_name=’STFullyConnected’, alg=<class ‘qsprpred.extra.gpu.model…eural_network.STFullyConnected’>, th=[0, 1, 10, 1100], random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
alg – Algorithm to use.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_2_STFullyConnected_REGRESSION_None(**kw)

Test the DNNModel model in one configuration [with _=’STFullyConnected_REGRESSION_None’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’STFullyConnected’, alg=<class ‘qsprpred.extra.gpu.model…eural_network.STFullyConnected’>, th=None, random_state=[None]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
alg – Algorithm to use.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_3_STFullyConnected_REGRESSION_1_42(**kw)

Test the DNNModel model in one configuration [with _=’STFullyConnected_REGRESSION_1_42’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’STFullyConnected’, alg=<class ‘qsprpred.extra.gpu.model…eural_network.STFullyConnected’>, th=None, random_state=[1, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
alg – Algorithm to use.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

testSingleTaskModel_4_STFullyConnected_REGRESSION_42_42(**kw)

Test the DNNModel model in one configuration [with _=’STFullyConnected_REGRESSION_42_42’, task=<TargetTasks.REGRESSION: ‘REGRESSION’>, alg_name=’STFullyConnected’, alg=<class ‘qsprpred.extra.gpu.model…eural_network.STFullyConnected’>, th=None, random_state=[42, 42]].

Parameters:

task – Task to test.
alg_name – Name of the algorithm.
alg – Algorithm to use.
th – Threshold to use for classification models.
random_state – Seed to be used for random operations.

validate_split(dataset): Check if the split has the data it should have after splitting.

class qsprpred.extra.gpu.models.tests.TestNNMonitoring(methodName='runTest')[source]

Bases: MonitorsCheckMixIn, TestCase

This class holds the tests for the monitoring classes.

Create an instance of the class that will use the named test method when executed. Raises a ValueError if the instance does not have a method with the specified name.

classmethod addClassCleanup(function, /, *args, **kwargs): Same as addCleanup, except the cleanup items are called even if setUpClass fails (unlike tearDownClass).

addCleanup(function, /, *args, **kwargs)

Add a function, with arguments, to be called when the test is completed. Functions added are called on a LIFO basis and are called after tearDown on test failure or success.

Cleanup items are called even if setUp fails (unlike tearDown).

addTypeEqualityFunc(typeobj, function)

Add a type specific assertEqual style function to compare a type.

This method is for use by TestCase subclasses that need to register their own type equality functions to provide nicer error messages.

Parameters:

typeobj – The data type to call this function on when both values are of the same type in assertEqual().
function – The callable taking two arguments and an optional msg= argument that raises self.failureException with a useful error message when the two arguments are not equal.

assertAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are unequal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is more than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

If the two objects compare equal then they will automatically compare almost equal.

assertCountEqual(first, second, msg=None)

Asserts that two iterables have the same elements, the same number of times, without regard to order.

self.assertEqual(Counter(list(first)),
Counter(list(second)))

Example:

[0, 1, 1] and [1, 0, 1] compare equal.

[0, 0, 1] and [0, 1] compare unequal.

assertDictEqual(d1, d2, msg=None)

assertEqual(first, second, msg=None): Fail if the two objects are unequal as determined by the ‘==’ operator.

assertFalse(expr, msg=None): Check that the expression is false.

assertGreater(a, b, msg=None): Just like self.assertTrue(a > b), but with a nicer default message.

assertGreaterEqual(a, b, msg=None): Just like self.assertTrue(a >= b), but with a nicer default message.

assertIn(member, container, msg=None): Just like self.assertTrue(a in b), but with a nicer default message.

assertIs(expr1, expr2, msg=None): Just like self.assertTrue(a is b), but with a nicer default message.

assertIsInstance(obj, cls, msg=None): Same as self.assertTrue(isinstance(obj, cls)), with a nicer default message.

assertIsNone(obj, msg=None): Same as self.assertTrue(obj is None), with a nicer default message.

assertIsNot(expr1, expr2, msg=None): Just like self.assertTrue(a is not b), but with a nicer default message.

assertIsNotNone(obj, msg=None): Included for symmetry with assertIsNone.

assertLess(a, b, msg=None): Just like self.assertTrue(a < b), but with a nicer default message.

assertLessEqual(a, b, msg=None): Just like self.assertTrue(a <= b), but with a nicer default message.

assertListEqual(list1, list2, msg=None)

A list-specific equality assertion.

Parameters:

list1 – The first list to compare.
list2 – The second list to compare.
msg – Optional message to use on failure instead of a list of differences.

assertLogs(logger=None, level=None)

Fail unless a log message of level level or higher is emitted on logger_name or its children. If omitted, level defaults to INFO and logger defaults to the root logger.

This method must be used as a context manager, and will yield a recording object with two attributes: output and records. At the end of the context manager, the output attribute will be a list of the matching formatted log messages and the records attribute will be a list of the corresponding LogRecord objects.

Example:

with self.assertLogs('foo', level='INFO') as cm:
    logging.getLogger('foo').info('first message')
    logging.getLogger('foo.bar').error('second message')
self.assertEqual(cm.output, ['INFO:foo:first message',
                             'ERROR:foo.bar:second message'])

assertMultiLineEqual(first, second, msg=None): Assert that two multi-line strings are equal.

assertNoLogs(logger=None, level=None)

Fail unless no log messages of level level or higher are emitted on logger_name or its children.

This method must be used as a context manager.

assertNotAlmostEqual(first, second, places=None, msg=None, delta=None)

Fail if the two objects are equal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is less than the given delta.

Note that decimal places (from zero) are usually not the same as significant digits (measured from the most significant digit).

Objects that are equal automatically fail.

assertNotEqual(first, second, msg=None): Fail if the two objects are equal as determined by the ‘!=’ operator.

assertNotIn(member, container, msg=None): Just like self.assertTrue(a not in b), but with a nicer default message.

assertNotIsInstance(obj, cls, msg=None): Included for symmetry with assertIsInstance.

assertNotRegex(text, unexpected_regex, msg=None): Fail the test if the text matches the regular expression.

assertRaises(expected_exception, *args, **kwargs)

Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments. If a different type of exception is raised, it will not be caught, and the test case will be deemed to have suffered an error, exactly as for an unexpected exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertRaises(SomeException):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertRaises is used as a context object.

The context manager keeps a reference to the exception as the ‘exception’ attribute. This allows you to inspect the exception after the assertion:

with self.assertRaises(SomeException) as cm:
    do_something()
the_exception = cm.exception
self.assertEqual(the_exception.error_code, 3)

assertRaisesRegex(expected_exception, expected_regex, *args, **kwargs)

Asserts that the message in a raised exception matches a regex.

Parameters:

expected_exception – Exception class expected to be raised.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertRaisesRegex is used as a context manager.

assertRegex(text, expected_regex, msg=None): Fail the test unless the text matches the regular expression.

assertSequenceEqual(seq1, seq2, msg=None, seq_type=None)

An equality assertion for ordered sequences (like lists and tuples).

For the purposes of this function, a valid ordered sequence type is one which can be indexed, has a length, and has an equality operator.

Parameters:

seq1 – The first sequence to compare.
seq2 – The second sequence to compare.
seq_type – The expected datatype of the sequences, or None if no datatype should be enforced.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual(set1, set2, msg=None)

A set-specific equality assertion.

Parameters:

set1 – The first set to compare.
set2 – The second set to compare.
msg – Optional message to use on failure instead of a list of differences.

assertSetEqual uses ducktyping to support different types of sets, and is optimized for sets specifically (parameters must support a difference method).

assertTrue(expr, msg=None): Check that the expression is true.

assertTupleEqual(tuple1, tuple2, msg=None)

A tuple-specific equality assertion.

Parameters:

tuple1 – The first tuple to compare.
tuple2 – The second tuple to compare.
msg – Optional message to use on failure instead of a list of differences.

assertWarns(expected_warning, *args, **kwargs)

Fail unless a warning of class warnClass is triggered by the callable when invoked with specified positional and keyword arguments. If a different type of warning is triggered, it will not be handled: depending on the other warning filtering rules in effect, it might be silenced, printed out, or raised as an exception.

If called with the callable and arguments omitted, will return a context object used like this:

with self.assertWarns(SomeWarning):
    do_something()

An optional keyword argument ‘msg’ can be provided when assertWarns is used as a context object.

The context manager keeps a reference to the first matching warning as the ‘warning’ attribute; similarly, the ‘filename’ and ‘lineno’ attributes give you information about the line of Python code from which the warning was triggered. This allows you to inspect the warning after the assertion:

with self.assertWarns(SomeWarning) as cm:
    do_something()
the_warning = cm.warning
self.assertEqual(the_warning.some_attribute, 147)

assertWarnsRegex(expected_warning, expected_regex, *args, **kwargs)

Asserts that the message in a triggered warning matches a regexp. Basic functioning is similar to assertWarns() with the addition that only warnings whose messages also match the regular expression are considered successful matches.

Parameters:

expected_warning – Warning class expected to be triggered.
expected_regex – Regex (re.Pattern object or string) expected to be found in error message.
args – Function to be called and extra positional args.
kwargs – Extra kwargs.
msg – Optional message used in case of failure. Can only be used when assertWarnsRegex is used as a context manager.

baseMonitorTest(monitor: BaseMonitor, monitor_type: Literal['hyperparam', 'crossval', 'test', 'fit'], neural_net: bool): Test the base monitor.

checkOptimization(model: QSPRModel, ds: QSPRDataset, optimizer: HyperparameterOptimization)

clearGenerated(): Remove the directories that are used for testing.

countTestCases()

createLargeMultitaskDataSet(name='QSPRDataset_multi_test', target_props=[{'name': 'HBD', 'task': <TargetTasks.MULTICLASS: 'MULTICLASS'>, 'th': [-1, 1, 2, 100]}, {'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
preparation_settings (dict) – dictionary containing preparation settings
random_state (int) – random state to use for splitting and shuffling

Returns:

a QSPRDataset object

Return type:

createLargeTestDataSet(name='QSPRDataset_test_large', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42, n_jobs=1, chunk_size=None)

Create a large dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createSmallTestDataSet(name='QSPRDataset_test_small', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], preparation_settings=None, random_state=42)

Create a small dataset for testing purposes.

Parameters:

name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
preparation_settings (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

createTestDataSetFromFrame(df, name='QSPRDataset_test', target_props=[{'name': 'CL', 'task': <TargetTasks.REGRESSION: 'REGRESSION'>}], random_state=None, prep=None, n_jobs=1, chunk_size=None)

Create a dataset for testing purposes from the given data frame.

Parameters:

df (pd.DataFrame) – data frame containing the dataset
name (str) – name of the dataset
target_props (List of dicts or TargetProperty) – list of target properties
random_state (int) – random state to use for splitting and shuffling
prep (dict) – dictionary containing preparation settings

Returns:

a QSPRDataset object

Return type:

debug(): Run the test without collecting errors in a TestResult

defaultTestResult()

classmethod doClassCleanups(): Execute all class cleanup functions. Normally called for you after tearDownClass.

doCleanups(): Execute all cleanup functions. Normally called for you after tearDown.

classmethod enterClassContext(cm): Same as enterContext, but class-wide.

enterContext(cm)

Enters the supplied context manager.

If successful, also adds its __exit__ method as a cleanup function and returns the result of the __enter__ method.

fail(msg=None): Fail immediately, with the given message.

failureException: alias of AssertionError

fileMonitorTest(monitor: FileMonitor, monitor_type: Literal['hyperparam', 'crossval', 'test', 'fit'], neural_net: bool): Test if the correct files are generated

fitTest(model: QSPRModel, ds: QSPRDataset)

Test model fitting, optimization and evaluation.

Parameters:

model (QSPRModel) – The model to test.
ds (QSPRDataset) – The dataset to use for testing.

classmethod getAllDescriptors()

Return a list of (ideally) all available descriptor sets. For now they need to be added manually to the list below.

TODO: would be nice to create the list automatically by implementing a descriptor set registry that would hold all installed descriptor sets.

Returns:: list of DescriptorCalculator objects
Return type:: list

getBigDF()

Get a large data frame for testing purposes.

Returns:: a pandas.DataFrame containing the dataset
Return type:: pd.DataFrame

classmethod getDataPrepGrid()

Return a list of many possible combinations of descriptor calculators, splits, feature standardizers, feature filters and data filters. Again, this is not exhaustive, but should cover a lot of cases.

Returns:: a generator that yields tuples of all possible combinations as stated above, each tuple is defined as: (descriptor_calculator, split, feature_standardizer, feature_filters, data_filters)
Return type:: grid

classmethod getDefaultCalculatorCombo()

Makes a list of default descriptor calculators that can be used in tests. It creates a calculator with only morgan fingerprints and rdkit descriptors, but also one with them both to test behaviour with multiple descriptor sets. Override this method if you want to test with other descriptor sets and calculator combinations.

Returns:: list of created DescriptorCalculator objects
Return type:: list

static getDefaultPrep(): Return a dictionary with default preparation settings.

getParamGrid(model: QSPRModel, grid: str) → dict

Get the parameter grid for a model.

Parameters:

model (QSPRModel) – The model to get the parameter grid for.
grid (str) – The grid type to get the parameter grid for.

Returns:

The parameter grid.

Return type: