alea.models package

Submodules

alea.models.blueice_extended_model module

class alea.models.blueice_extended_model.BlueiceExtendedModel(parameter_definition: dict, likelihood_config: dict, **kwargs)[source]

Bases: StatisticalModel

A statistical model based on blueice likelihoods.

This class extends the StatisticalModel class and provides methods for generating data and computing likelihoods based on blueice.

parameters

Parameters object containing the parameters of the model.

Type: Parameters

data

Data of the statistical model.

Type: dict

is_data_set

Whether data is set.

Type: bool

_likelihood

A blueice LogLikelihoodSum instance.

Type: LogLikelihoodSum

likelihood_names

List of likelihood names.

Type: list

livetime_parameter_names

List of the name of the livetime of each term, None if not specified

Type: list

data_generators

List of data generators for each likelihood term.

Type: list

Parameters

parameter_definition (dict) – A dictionary defining the model parameters.
likelihood_config (dict) – A dictionary defining the likelihood.

Todo

analysis_space could be inferred from the data (assert that all sources have the same analysis_space)

__init__(parameter_definition: dict, likelihood_config: dict, **kwargs)[source]

Initialize the statistical model from parameter and likelihood config dicts.

Parameters

parameter_definition (dict) – A dictionary defining the model parameters.
likelihood_config (dict) – A dictionary defining the likelihood.

_build_data_generators() → list[source]

Build data generators for all likelihood terms.

Returns: List of data generators for each likelihood term.
Return type: list

Todo

Also implement data generator for ancillary ll term.

_build_ll_from_config(likelihood_config: dict, template_path: Optional[str] = None) → LogLikelihoodSum[source]

Iterate through all likelihood terms and build blueice likelihood instances.

Parameters: likelihood_config (dict) – A dictionary defining the likelihood.
Returns: A blueice LogLikelihoodSum instance.
Return type: LogLikelihoodSum

_generate_ancillary(**generate_values) → dict[source]

Generate data for the ancillary likelihood.

Keyword Arguments: generate_values (dict) – A dictionary of parameter values.
Returns: A numpy structured array of ancillary measurements.
Return type: numpy.array

_generate_data(**generate_values) → dict[source]

Generate data for all likelihood terms and ancillary likelihood.

Keyword Arguments: generate_values (dict) – A dictionary of parameter values.
Returns: A dict of data-sets, with key of the likelihood term name, “ancillary” and “generate_values”.
Return type: dict

_generate_science_data(**generate_values) → dict[source]: Generate the science data for all likelihood terms except the ancillary likelihood.

_process_blueice_config(config, template_folder_list)[source]: Process the blueice config from config.

_set_default_ptype()[source]

Check if all parameters have a ptype that is in the list of allowed ptypes.

If no ptype is specified, set the default ptype “needs_reinit”.

_set_efficiency(source: dict, ll)[source]

Set the efficiency of a source in the blueice ll.

Parameters

source (dict) – A dictionary defining the source.
ll (LogLikelihood) – A blueice LogLikelihood instance.

Raises

ValueError – If the efficiency_name is not specified in the source.

property all_source_names: list: Return a sorted list of all source names across all likelihood terms.

property data: Union[dict, list]: Return the data of the statistical model.

classmethod from_config(config_file_path: str, **kwargs) → BlueiceExtendedModel[source]

Initialize the statistical model from a yaml config file.

Parameters: config_file_path (str) – Path to the yaml config file.
Returns: Statistical model.
Return type: BlueiceExtendedModel

get_expectation_values(per_likelihood_term=False, **kwargs) → dict[source]

Return total expectation values summed over all likelihood terms with the same name.

Parameters

per_likelihood_term (bool) – If True, return expectation values per likelihood term. Otherwise, sum each source over all likelihood terms.
kwargs – Named parameters

Returns

Dictionary of expectation values. If per_likelihood_term is True, the dictionary: has the form {likelihood_name: {source_name: expectation_value, …}, …}.

Return type

dict

Todo

Make a self.likelihood_terms dict with the likelihood names as keys and the corresponding likelihood terms as values.

get_source_histograms(likelihood_name: str, expected_events=False, **kwargs) → dict[source]

Return the pdfs or histograms of all sources for a given likelihood term.

Parameters

likelihood_name (str) – Name of the likelihood term.
expected_events (bool) – If True, return the histograms containing the number of expected events.
kwargs – Named parameters.

Returns

Dictionary containing a multihist object for each source.

Return type

dict

get_source_name_list(likelihood_name: str) → list[source]

Return a list of source names for a given likelihood term.

The order matches the source column of the data, so this can be used to map indices in the data to a source name.

Parameters: likelihood_name (str) – Name of the likelihood.
Returns: List of source names.
Return type: list

property likelihood_list: List: Return a list of likelihood terms.

property likelihood_parameters: List: Return a list of likelihood parameters.

store_data(file_name, data_list, data_name_list=None, metadata=None)[source]

Store data in a file.

Append the generate_values to the data_name_list.

store_real_data(file_name: str, real_data_list: list, metadata=None)[source]

Store real data in a file with toydata format.

Parameters

file_name (str) – Name of the file.
real_data_list (list) – List of np.array of real data.

class alea.models.blueice_extended_model.CustomAncillaryLikelihood(parameters: Parameters)[source]

Bases: LogAncillaryLikelihood

Ancillary likelihood providing constraint terms for nuisance parameters.

Can be used to add constraint terms for parameters of the likelihood.

parameters

Parameters object containing the parameters to be constrained.

Type: Parameters

constraint_functions

Dict of constraint functions for all ancillary parameters.

Type: dict

__init__(parameters: Parameters)[source]: Initialize the CustomAncillaryLikelihood.

_get_constraint_functions(**generate_values) → dict[source]

Get callable constraint functions for all ancillary parameters.

Keyword Arguments: generate_values (dict) – A dictionary of parameter values.
Returns: Dict of constraint functions for all ancillary parameters.
Return type: dict

Todo

Implement str-type uncertainties.

ancillary_sum(evaluate_at: dict) → float[source]

Return the sum of all constraint terms.

Parameters: evaluate_at (dict) – Values of the ancillary measurements.
Returns: Sum of all constraint terms.
Return type: float

property constraint_terms: dict

Return a dict of all constraint terms (logpdf callables) of the ancillary likelihood.

Returns: Dict of all constraint terms function.
Return type: dict

set_data(d: array)[source]

Set the data of the ancillary likelihood (ancillary measurements).

Parameters: d (numpy.array) – Data of ancillary measurements, stored as numpy array.

alea.models package

Submodules

alea.models.blueice_extended_model module

Module contents