pcntoolkit.normative

Module providing entry points for fitting and predicting with normative models from the command line.

Functions

entrypoint(*args)
fit(→ None)	Fit a new normative model.
fit_predict(→ None)	Fit a normative model and predict response variables.
get_argparser(→ argparse.ArgumentParser)	Get an argument parser for the normative modeling functions.
get_conf_dict_from_args(→ dict[str, ...)	Parse the arguments and return a dictionary with the configuration options.
load_data(→ pcntoolkit.dataio.norm_data.NormData)	Load the data from the configuration dictionary.
load_response_vars(→ tuple[numpy.ndarray, ...)	Load response variables from file. This will load the data and mask it if
load_test_data(→ pcntoolkit.dataio.norm_data.NormData)	Load the test data from the file specified in the configuration dictionary.
main(→ None)	Main function to run the normative modeling functions.
make_synthetic_data(→ None)	Create synthetic data for testing.
predict(→ None)	Predict response variables using a saved normative model.

Module Contents

entrypoint(*args)

fit(conf_dict: dict) → None

Fit a new normative model.

Parameters:

conf_dict – Dictionary containing configuration options

fit_predict(conf_dict: dict) → None

Fit a normative model and predict response variables.

Parameters:

conf_dict – Dictionary containing configuration options

get_argparser() → argparse.ArgumentParser

Get an argument parser for the normative modeling functions.

Returns:

argparse.ArgumentParser: The argument parser

get_conf_dict_from_args() → dict[str, str | int | float | bool]

Parse the arguments and return a dictionary with the configuration options.

Raises:

ValueError: Raised if an argument is specified twice.

Returns:

dict[str, str | int | float | bool]: A dictionary with the configuration option, parsed to the correct type.

load_data(conf_dict: dict) → pcntoolkit.dataio.norm_data.NormData

Load the data from the configuration dictionary.

Returns:

NormData: NormData object containing the data

load_response_vars(datafile: str, maskfile: str | None = None, vol: bool = True) → tuple[numpy.ndarray, numpy.ndarray | None]

Load response variables from file. This will load the data and mask it if necessary. If the data is in ascii format it will be converted into a numpy array. If the data is in neuroimaging format it will be reshaped into a 2D array (observations x variables) and a mask will be created if necessary.

Parameters:

• datafile – File containing the response variables

• maskfile – Mask file (nifti only)

• vol – If True, load the data as a 4D volume (nifti only)

Returns Y:

Response variables

Returns volmask:

Mask file (nifti only)

load_test_data(conf_dict: dict) → pcntoolkit.dataio.norm_data.NormData

Load the test data from the file specified in the configuration dictionary.

Args:

conf_dict (dict): dictionary containing the configuration options

Returns:

NormData: NormData object containing the test data

main(*args) → None

Main function to run the normative modeling functions.

Raises:

ValueError: If the function specified in the configuration dictionary is unknown.

make_synthetic_data() → None

Create synthetic data for testing.

predict(conf_dict: dict) → None

Predict response variables using a saved normative model.

Parameters:

conf_dict – Dictionary containing configuration options