bluepyemodel.access_point.local

LocalAccessPoint class.

Classes

LocalAccessPoint(emodel[, emodel_dir, ...])

Access point to access configuration files and e-models when stored locally.
class LocalAccessPoint(emodel, emodel_dir=None, etype=None, ttype=None, mtype=None, species=None, brain_region=None, iteration_tag=None, synapse_class=None, final_path=None, recipes_path=None, legacy_dir_structure=False, with_seeds=False)

Bases: DataAccessPoint

Access point to access configuration files and e-models when stored locally.

Init

Parameters:

  • emodel (str) – name of the emodel.

  • emodel_dir (str) – path to the working directory. Default to current working directory. If iteration_tag is not None, it will be ./run/iteration_tag

  • ttype (str) – name of the t-type

  • iteration_tag (str) – iteration tag

  • final_path (str) – path to final.json which will contain the optimised models.

  • recipes_path (str) –

    path to the json file which should contain the path to the configuration files used for each etype. The content of this file should follow the format:

    {
    "emodelname": {
        "morph_path": "./morphologies/",
        "morphology": [[morphologyname", "morphfile.asc"]],
        "params": "config/params/pyr.json",
        "protocol": "config/protocols/emodelname.json",
        "features": "config/features/emodelname.json"
    }
}

  • legacy_dir_structure (bool) – if true, the path coming from the recipes will be replaced by “self.emodel_dir / emodel / json_path_from_recipes”. To be deprecated.

  • with_seed (bool) – allows for emodel_seed type of emodel names in final.json

classmethod add_entry_recipes(recipes_path, emodel, morph_path, morphology, parameters_path, protocols_path, features_path, pipeline_settings=None)

Append an entry to the recipes, create the file if it does not exist

Parameters:

  • recipes_path (str) – path to the json containig the recipes.

  • emodel (str) – name of the emodel.

  • morph_path (str) – path the directory containing the morphologies.

  • morphology (list of str) – name of the morphology and name of the file.

  • parameters_path (str) – path to the json that contains the parameters.

  • protocols_path (str) – path to the json that contains the protocols.

  • features_path (str) – path to the json that contains the features.

  • pipeline_settings (dict) – pipeline settings.

check_emodel_workflow_configurations(emodel_workflow)

Emodel workflow is not used in local, so always return True to let the workflow run

create_emodel_workflow(state='not launched')

Create an empty EModelWorkflow instance. EModel workflow should not be used in local

format_emodel_data(model_data)

Format emodel data.

get_all_recipes()

Load the recipes from a json file.

See docstring of __init__ for the format of the file of recipes.

get_available_mechanisms()

Get the list of names of the available mechanisms

get_available_morphologies()

Get the list of names of available morphologies

get_emodel(lock_file=True)

Get dict with parameter of single emodel (including seed if any)

get_emodel_names()

Get the list of all the names of emodels

Returns:

keys are emodel names with seed, values are names without seed.

Return type:

dict

get_emodel_workflow()

Emodel workflow is not used in local, so return None here

get_emodels(emodels=None)

Get a list of emodels

Parameters:

emodels (list) – list of names of the emodels.

get_final(lock_file=True)

Get emodel dictionary from final.json.

get_final_content(lock_file=True)

Return the final content from recipes if available, or fallback to final.json

get_fitness_calculator_configuration(record_ions_and_currents=False)

Get the configuration of the fitness calculator (efeatures and protocols)

get_json(recipe_entry)

Helper function to load a json using path in recipe.

get_json_path(recipe_entry)

Helper function to get a json path in recipe.

get_mechanisms_directory()

Return the path to the directory containing the mechanisms for the current emodel

get_model_configuration(skip_get_available_morph=False)

Get the configuration of the model, including parameters, mechanisms and distributions

Parameters:

skip_get_available_morph (bool) – only used in nexus access point.

get_model_name_for_final(seed)

Return model name used as key in final.json.

get_morphologies()

Get the name and path to the morphologies from the recipes.

Returns:

[{‘name’: morph_name, ‘path’: ‘morph_path’}]. Might contain the additional entries “seclist_names” and “secarray_names” if they are present in the recipes.

Return type:

morphology_definition (list)

get_recipes()

Load the recipes from a json file for an emodel.

See docstring of __init__ for the format of the file of recipes.

get_targets_configuration()

Get the configuration of the targets (targets and ephys files used)

has_best_model(seed)

Check if the best model has been stored.

has_fitness_calculator_configuration()

Check if the fitness calculator configuration exists

has_model_configuration()

Check if the model configuration exists

has_pipeline_settings()

Returns True if pipeline settings are present in the recipes

has_targets_configuration()

Check if the target configuration exists

is_checked_by_validation(seed)

Check if the emodel with a given seed has been checked by Validation task.

is_validated()

Check if enough models have been validated.

property morph_dir

Return the morphology directory as read from the recipes, or fallback to ‘morphology’.

property morph_path

Return the path to the morphology file as read from the recipes.

save_final(final, final_path, lock_file=True)

Save final emodels in json

set_emodel(emodel)

Setter for the name of the emodel, check it exists (with or without seed) in recipe.

set_unfrozen_params(params)

Freeze parameters for partial optimisation.

sonata_exists(seed)

Returns True if the sonata hoc file has been exported

store_emodel(emodel)

Store an emodel obtained from BluePyOpt in the final.json. Note that if a model in the final.json has the same key (emodel__iteration_tag__seed), it will be overwritten.

store_emodels_hoc(only_validated=False, only_best=True, seeds=None, map_function=<class 'map'>, new_emodel_name=None, description=None)

Not Implemented

store_emodels_sonata(only_validated=False, only_best=True, seeds=None, map_function=<class 'map'>, new_emodel_name=None, description=None)

Not Implemented

store_fitness_calculator_configuration(configuration)

Store a fitness calculator configuration

store_hocs(only_validated=False, only_best=True, seeds=None, map_function=<class 'map'>, new_emodel_name=None, description=None, output_base_dir='export_emodels_hoc')

Not Implemented

store_model_configuration(configuration, path=None)

Store a model configuration as a json

store_or_update_emodel(emodel)

Calls store_emodel.

store_or_update_emodel_workflow(emodel_workflow)

Emodel workflow is not used in local, so pass

store_targets_configuration(configuration)

Store the configuration of the targets (targets and ephys files used)