neuronunit.optimization package

Submodules

neuronunit.optimization.algorithms module

Optimisation class Copyright (c) 2016, EPFL/Blue Brain Project

This file is part of BluePyOpt <https://github.com/BlueBrain/BluePyOpt>

This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License version 3.0 as published by the Free Software Foundation.

This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

neuronunit.optimization.algorithms.eaAlphaMuPlusLambdaCheckpoint(population, toolbox, mu, cxpb, mutpb, ngen, stats=None, halloffame=None, pf=None, nelite=3, cp_frequency=1, cp_filename=None, continue_cp=False, selection='selNSGA2', td=None)

neuronunit.optimization.algorithms.gene_bad(offspring)

neuronunit.optimization.bp_opt module

Optimisation class

class neuronunit.optimization.bp_opt.SciUnitOptimization(error_criterion=None, evaluator=None, selection='selIBEA', benchmark=False, seed=1, offspring_size=15, elite_size=3, eta=10, mutpb=1.0, cxpb=1.0, map_function=None, backend=None, nparams=10, provided_dict={})

Bases: bluepyopt.optimisations.Optimisation

DEAP Optimisation class

grid_sample_init(nparams)

run(max_ngen=25, offspring_size=None, continue_cp=False, cp_filename=None, cp_frequency=0)

Run optimisation

set_evaluate()

set_pop()

setnparams(nparams=10, provided_dict=None)

setup_deap()

Set up optimisation

transdict(dictionaries)

class neuronunit.optimization.bp_opt.WSFloatIndividual(*args, **kwargs)

Bases: float

Individual consisting of list with weighted sum field

set_fitness(obj_size)

class neuronunit.optimization.bp_opt.WSListIndividual(*args, **kwargs)

Bases: list

Individual consisting of list with weighted sum field

set_fitness(obj_size)

class neuronunit.optimization.bp_opt.WeightedSumFitness(values=(), obj_size=None)

Bases: deap.base.Fitness

Fitness that compares by weighted sum

norm

Frobenius norm of values

sum

Weighted sum of values

weighted_sum

Weighted sum of wvalues

neuronunit.optimization.covariance_adaption_approach module

neuronunit.optimization.data_transport_container module

class neuronunit.optimization.data_transport_container.DataTC

Bases: object

Data Transport Container

This Object class serves as a data type for storing rheobase search attributes and apriori model parameters, with the distinction that unlike the NEURON model this class can be cheaply transported across HOSTS/CPUs

add_constant()

get_ss

neuronunit.optimization.exhaustive_search module

class neuronunit.optimization.exhaustive_search.WSListIndividual(*args, **kwargs)

Bases: list

Individual consisting of list with weighted sum field

neuronunit.optimization.exhaustive_search.add_constant(hold_constant, pop)

neuronunit.optimization.exhaustive_search.build_chunk_grid(npoints, free_params, hold_constant=None, mp_in=None)

neuronunit.optimization.exhaustive_search.chunks(l, n)

neuronunit.optimization.exhaustive_search.create_a_map

neuronunit.optimization.exhaustive_search.create_grid(mp_in=None, npoints=3, free_params=None, ga=None)

check for overlap in parameter space.

neuronunit.optimization.exhaustive_search.reduce_params

neuronunit.optimization.exhaustive_search.run_grid(npoints, tests, provided_keys=None, hold_constant=None, ranges=None)

neuronunit.optimization.exhaustive_search.run_rick_grid(rick_grid, tests, td)

neuronunit.optimization.exhaustive_search.run_simple_grid(npoints, tests, ranges, free_params, hold_constant=None)

neuronunit.optimization.exhaustive_search.sample_points(iter_dict, npoints=3)

neuronunit.optimization.exhaustive_search.tfc2i(x, y, z, err)

translate_float_cube_to_index

Takes x, y, z values as lists and returns a 2D numpy array

neuronunit.optimization.exhaustive_search.tfg2i

translate_float_grid_to_index Takes x, y, z values as lists and returns a 2D numpy array

neuronunit.optimization.exhaustive_search.transdict

neuronunit.optimization.exhaustive_search.update_dtc_grid

neuronunit.optimization.get_neab module

neuronunit.optimization.get_neab.get_neuron_criteria(cell_id, file_name=None)

neuronunit.optimization.get_neab.get_obs(pipe)

neuronunit.optimization.get_neab.get_tests()

neuronunit.optimization.get_neab.neuroelectro_summary_observation(neuron, ontology)

neuronunit.optimization.get_neab.replace_zero_std(electro_tests)

neuronunit.optimization.get_neab.substitute_criteria(observations_donar, observations_acceptor)

neuronunit.optimization.get_neab.substitute_parallel_for_serial(electro_tests)

neuronunit.optimization.get_neab.update_amplitude(test, tests, score)

neuronunit.optimization.model_parameters module

neuronunit.optimization.model_parameters.transcribe_units(input_dic)

Move between OSB unit conventions and NEURON unit conventions.

neuronunit.optimization.model_parameters.type2007 = {'CH': (50, 1.5, -60, -40, 25, 0.03, 1, -40, 150, 3), 'FS': (20, 1.0, -55, -40, 25, 0.2, -2, -45, -55, 5), 'IB': (150, 1.2, -75, -45, 50, 0.01, 5, -56, 130, 2), 'LTS': (100, 1.0, -56, -42, 40, 0.03, 8, -53, 20, 4), 'RS': (100, 0.7, -60, -40, 35, 0.03, -2, -50, 100, 1), 'RTN': (40, 0.25, -65, -45, 0, 0.015, 10, -55, 50, 7), 'RTN_burst': (40, 0.25, -65, -45, 0, 0.015, 10, -55, 50, 7), 'TC': (200, 1.6, -60, -50, 35, 0.01, 15, -60, 10, 6), 'TC_burst': (200, 1.6, -60, -50, 35, 0.01, 15, -60, 10, 6)}

temp = {k:[] for k in [‘C’,’k’,’vr’,’vt’,’vpeak’,’a’,’b’,’c’,’d’] } for i,k in enumerate(temp.keys()):

for v in type2007.values():

temp[k].append(v[i])

explore_param = {k:(np.min(v),np.max(v)) for k,v in temp.items()} model_params = OrderedDict(explore_param)

neuronunit.optimization.opt_man module

neuronunit.optimization.optimization_management module

class neuronunit.optimization.optimization_management.WSFloatIndividual(*args, **kwargs)

Bases: float

Individual consisting of list with weighted sum field

class neuronunit.optimization.optimization_management.WSListIndividual(*args, **kwargs)

Bases: list

Individual consisting of list with weighted sum field

neuronunit.optimization.optimization_management.active_values(keyed, rheobase)

neuronunit.optimization.optimization_management.add_constant(hold_constant, pop, td)

neuronunit.optimization.optimization_management.allocate_worst(dtc, tests)

neuronunit.optimization.optimization_management.bridge_judge(test_and_models)

neuronunit.optimization.optimization_management.create_subset(nparams=10, boundary_dict=None)

neuronunit.optimization.optimization_management.dtc_to_rheo(dtc)

neuronunit.optimization.optimization_management.evaluate(dtc)

neuronunit.optimization.optimization_management.filtered(pop, dtcpop)

neuronunit.optimization.optimization_management.format_test(dtc)

neuronunit.optimization.optimization_management.get_rh(dtc, rtest)

neuronunit.optimization.optimization_management.get_trans_list(param_dict)

neuronunit.optimization.optimization_management.init_pop(pop, td, tests)

neuronunit.optimization.optimization_management.make_up_lost(pop, dtcpop, td)

neuronunit.optimization.optimization_management.mint_generic_model(backend)

neuronunit.optimization.optimization_management.new_genes(pop, dtcpop, td)

some times genes explored will not return un-usable simulation parameters genes who have no rheobase score will be discarded.

BluePyOpt needs a stable gene number however

This method finds how many genes have been discarded, and tries to build new genes from the existing distribution of gene values, by mimicing a normal random distribution of genes that are not deleted. if a rheobase value cannot be found for a given set of dtc model more_attributes delete that model, or rather, filter it out above, and make new genes based on the statistics of remaining genes. it’s possible that they wont be good models either, so keep trying in that event. a new model from the mean of the pre-existing model attributes.

neuronunit.optimization.optimization_management.nunit_evaluation(dtc)

neuronunit.optimization.optimization_management.nunit_evaluation_df(dtc)

neuronunit.optimization.optimization_management.obtain_rheobase(pop, td, tests)

Calculate rheobase for a given population pop Ordered parameter dictionary td and rheobase test rt

neuronunit.optimization.optimization_management.parallel_route(pop, dtcpop, tests, td)

neuronunit.optimization.optimization_management.passive_values(keyed)

neuronunit.optimization.optimization_management.run_ga(explore_edges, max_ngen, test, free_params=None, hc=None, NSGA=None, MU=None, seed_pop=None, model_type='RAW')

neuronunit.optimization.optimization_management.score_proc(dtc, t, score)

neuronunit.optimization.optimization_management.serial_route(pop, td, tests)

parallel list mapping only works with an iterable collection. Serial route is intended for single items.

neuronunit.optimization.optimization_management.switch_logic(tests)

neuronunit.optimization.optimization_management.test_runner(pop, td, tests, single_spike=True)

neuronunit.optimization.optimization_management.transform(xargs)

neuronunit.optimization.optimization_management.update_deap_pop(pop, tests, td, backend=None, hc=None)

Inputs a population of genes (pop). Returned neuronunit scored DataTransportContainers (dtcpop). This method converts a population of genes to a population of Data Transport Containers, Which act as communicatable data types for storing model attributes. Rheobase values are found on the DTCs DTCs for which a rheobase value of x (pA)<=0 are filtered out DTCs are then scored by neuronunit, using neuronunit models that act in place.

neuronunit.optimization.optimization_management.update_dtc_pop(pop, td)

inputs a population of genes/alleles, the population size MU, and an optional argument of a rheobase value guess outputs a population of genes/alleles, a population of individual object shells, ie a pickleable container for gene attributes. Rationale, not every gene value will result in a model for which rheobase is found, in which case that gene is discarded, however to compensate for losses in gene population size, more gene samples must be tested for a successful return from a rheobase search. If the tests return are successful these new sampled individuals are appended to the population, and then their attributes are mapped onto corresponding virtual model objects.

neuronunit.optimization.optimization_management.write_opt_to_nml

Write optimimal simulation parameters back to NeuroML.

neuronunit.optimization.results_analysis module

neuronunit.optimization.results_analysis.error_domination(dtc_ga, dtc_grid)

neuronunit.optimization.results_analysis.make_report(grid_results, ga_out, nparams, pop=None)

neuronunit.optimization.results_analysis.min_max(pop)

neuronunit.optimization.results_analysis.param_distance(dtc_ga_attrs, dtc_grid_attrs, td)

Module contents

NeuronUnit Optimization classes