bmtk.simulator.core package

Subpackages

• bmtk.simulator.core.sonata_reader package
  • Submodules
  • bmtk.simulator.core.sonata_reader.edge_adaptor module
  • bmtk.simulator.core.sonata_reader.network_reader module
  • bmtk.simulator.core.sonata_reader.node_adaptor module
  • Module contents

Submodules

bmtk.simulator.core.edge_population module

class bmtk.simulator.core.edge_population.EdgePopulation

Bases: object

initialize(network)

source_nodes

target_nodes

class bmtk.simulator.core.edge_population.SimEdge

Bases: object

gid

node_id

bmtk.simulator.core.graph module

class bmtk.simulator.core.graph.SimGraph

Bases: object

add_component(key, value)

Add a component key-value pair

Parameters:

• key – name of component
• value – value

add_edges(sonata_file, populations=None, source_pop=None, target_pop=None)

Parameters:

• sonata_file –
• populations –
• source_pop –
• target_pop –

Returns:

add_nodes(sonata_file, populations=None)

Add nodes from a network to the graph.

Parameters:

• sonata_file – A NodesFormat type object containing list of nodes.
• populations – name/identifier of network. If none will attempt to retrieve from nodes object

build_nodes()

build_recurrent_edges()

external_edge_populations(src_pop, trg_pop)

classmethod from_config(conf, **properties)

Generates a graph structure from a json config file or dictionary.

Parameters:

• conf – name of json config file, or a dictionary with config parameters
• properties – optional properties.

Returns:

A graph object of type cls

get_component(key)

Get the value of item in the components dictionary.

Parameters:key – name of component Returns:value assigned to component

get_node_population(population_name)

get_node_populations()

get_node_set(node_set)

io

model_type_col = 'model_type'

node_populations

bmtk.simulator.core.io_tools module

class bmtk.simulator.core.io_tools.IOUtils

Bases: object

For logging/mkdir commands we sometimes need to use different MPI classes depending on the simulator being used (NEST and NEURON have their own barrier functions that don’t work well with mpi). We also need to be able to adjust the logging levels/format at run-time depending on the simulator/configuration options.

Thus the bulk of the io and logging functions are put into their own class and can be overwritten by specific simulator modules

barrier()

MPI Barrier call

log_error(message)

log_exception(message, raise_exception=True)

log_info(message, all_ranks=False)

log_to_console

log_warning(message, all_ranks=False, display_once=False)

logger

quiet_simulator()

Turns off logging/messages of the native simulator

set_log_format(format_str)

set_log_level(loglevel)

setup_output_dir(output_dir, log_file, overwrite=True)

bmtk.simulator.core.network_reader module

class bmtk.simulator.core.network_reader.EdgesReader

Bases: object

initialize(network)

mixed = 2

mixed_connections

recurrent = 0

recurrent_connections

set_connection_type(src_pop, trg_pop)

source_nodes

target_nodes

unknown = 0

virtual = 1

virtual_connections

class bmtk.simulator.core.network_reader.NodesReader

Bases: object

initialize(network)

internal_nodes_only

classmethod load(**properties)

mixed_nodes

name

virtual_nodes_only

bmtk.simulator.core.node_population module

class bmtk.simulator.core.node_population.NodePopulation

Bases: object

initialize(network)

internal_nodes_only

classmethod load(**properties)

mixed_nodes

name

virtual_nodes_only

class bmtk.simulator.core.node_population.SimNode

Bases: object

gid

node_id

bmtk.simulator.core.node_sets module

class bmtk.simulator.core.node_sets.NodeSet(filter_params, network)

Bases: object

gids()

nodes()

population_names()

populations()

class bmtk.simulator.core.node_sets.NodeSetAll(network)

Bases: bmtk.simulator.core.node_sets.NodeSet

bmtk.simulator.core.node_sets.to_list(val)

bmtk.simulator.core.pyfunction_cache module

bmtk.simulator.core.pyfunction_cache.add_cell_model(func, directive, model_type='*', overwrite=True)

bmtk.simulator.core.pyfunction_cache.add_cell_processor(func, name=None, overwrite=True)

bmtk.simulator.core.pyfunction_cache.add_synapse_model(func, name=None, overwrite=True)

bmtk.simulator.core.pyfunction_cache.add_weight_function(func, name=None, overwrite=True)

bmtk.simulator.core.pyfunction_cache.cell_model(*wargs, **wkwargs)

A decorator for registering NEURON cell loader functions.

bmtk.simulator.core.pyfunction_cache.load_py_modules(cell_models=None, syn_models=None, syn_weights=None, cell_processors=None)

bmtk.simulator.core.pyfunction_cache.synapse_model(*wargs, **wkwargs)

A decorator for registering NEURON synapse loader functions.

bmtk.simulator.core.pyfunction_cache.synaptic_weight(*wargs, **wkwargs)

A decorator for registering a function as a synaptic weight function. To use either:

@synaptic_weight
def weight_function():
    ...

or:

@synaptic_weight(name='name_in_edge_types')
def weight_function():
    ...

Once the decorator has been attached and imported the functions will automatically be added to py_modules and BMTK will when assigning synaptic/gap junction weights for edges with matching “weight_function” attribute

bmtk.simulator.core.simulation_config module

class bmtk.simulator.core.simulation_config.SimulationConfig(*args, **kwargs)

Bases: bmtk.utils.sonata.config.sonata_config.SonataConfig

A special version of SonataConfig that contains some methods that can be used by the simulators. Mainly it contains the build_env() method which can be called to setup logging plus initialize the ‘output_dir’ folder.

Functionality that is specific to SONATA should go in SonataConfig

build_env(force=False)

Creates the folder(s) set in ‘output’ section, sets up logging and copies over the configuration

copy_to_output()

io

validate()

validator

bmtk.simulator.core.simulation_config.from_dict(config_dict, validator=None, **opts)

bmtk.simulator.core.simulation_config.from_json(config_file, validator=None, **opts)

bmtk.simulator.core.simulation_config_validator module

class bmtk.simulator.core.simulation_config_validator.SimulationConfigValidator(schema, types=(), resolver=None, format_checker=None, file_formats=())

Bases: jsonschema.validators.Draft4Validator

A JSON Schema validator class that will store a schema (passed into the constructor) and validate a json file. It has all the functionality of the JSONSchema format, plus includes special types and parameters like making sure a value is a file or directory type, checking csv files, etc.

To Use:

validator = SimConfigValidator(json_schema.json) validator.validate(file.json)

is_type(instance, dtype)

bmtk.simulator.core.simulator module

class bmtk.simulator.core.simulator.Simulator

Bases: object

add_mod(module)

run()

bmtk.simulator.core.simulator_network module

class bmtk.simulator.core.simulator_network.SimNetwork

Bases: object

add_component(name, path)

add_edges(edge_population)

add_node_set(name, node_set)

add_nodes(node_population)

build()

build_nodes()

build_recurrent_edges(**opts)

build_virtual_connections()

classmethod from_builder(network)

classmethod from_config(conf, **properties)

Generates a graph structure from a json config file or dictionary.

Parameters:

• conf – name of json config file, or a dictionary with config parameters
• properties – optional properties.

Returns:

A graph object of type cls

classmethod from_manifest(manifest_json)

get_component(name)

get_edge_adaptor(name)

get_node_adaptor(name)

get_node_groups(populations=None)

get_node_population(name)

get_node_populations()

get_node_set(node_set)

get_node_sets(populations=None, groupby=None, **filterby)

has_component(name)

io

load_gap_junc_files(gj_dic)

node_populations

node_properties(populations=None)

py_function_caches

recurrent_edges

Module contents