class SimulatorAdapter(ABCAsynchronous):
"""
Interface between the Simulator and the Framework.
"""
_ui_name = "Simulation Core"
algorithm = None
available_models = get_traited_subclasses(Model)
available_monitors = get_traited_subclasses(Monitor)
available_integrators = get_traited_subclasses(Integrator)
available_couplings = get_traited_subclasses(Coupling)
# This is a list with the monitors that actually return multi dimensions for the state variable dimension.
# We exclude from this for example EEG, MEG or Bold which return
HAVE_STATE_VARIABLES = ["GlobalAverage", "SpatialAverage", "Raw", "SubSample", "TemporalAverage"]
def __init__(self):
super(SimulatorAdapter, self).__init__()
self.log.debug("%s: Initialized..." % str(self))
def get_input_tree2(self):
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface_experimental
return result
def get_input_tree(self):
"""
Return a list of lists describing the interface to the simulator. This
is used by the GUI to generate the menus and fields necessary for
defining a simulation.
"""
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface[self.INTERFACE_ATTRIBUTES]
# We should add as hidden the Simulator State attribute.
result.append({self.KEY_NAME: 'simulation_state',
self.KEY_TYPE: 'tvb.datatypes.simulation_state.SimulationState',
self.KEY_LABEL: "Continuation of", self.KEY_REQUIRED: False, self.KEY_UI_HIDE: True})
return result
def get_output(self):
"""
:returns: list of classes for possible results of the Simulator.
"""
return [time_series.TimeSeries]
def configure(self, model, model_parameters, integrator, integrator_parameters, connectivity,
monitors, monitors_parameters=None, surface=None, surface_parameters=None, stimulus=None,
coupling=None, coupling_parameters=None, initial_conditions=None,
conduction_speed=None, simulation_length=0, simulation_state=None):
"""
Make preparations for the adapter launch.
"""
self.log.debug("available_couplings: %s..." % str(self.available_couplings))
self.log.debug("coupling: %s..." % str(coupling))
self.log.debug("coupling_parameters: %s..." % str(coupling_parameters))
self.log.debug("%s: Initializing Model..." % str(self))
noise_framework.build_noise(model_parameters)
model_instance = self.available_models[str(model)](**model_parameters)
self._validate_model_parameters(model_instance, connectivity, surface)
self.log.debug("%s: Initializing Integration scheme..." % str(self))
noise_framework.build_noise(integrator_parameters)
integr = self.available_integrators[integrator](**integrator_parameters)
self.log.debug("%s: Instantiating Monitors..." % str(self))
monitors_list = []
for monitor_name in monitors:
if (monitors_parameters is not None) and (str(monitor_name) in monitors_parameters):
current_monitor_parameters = monitors_parameters[str(monitor_name)]
HRFKernelEquation.build_equation_from_dict('hrf_kernel', current_monitor_parameters, True)
monitors_list.append(self.available_monitors[str(monitor_name)](**current_monitor_parameters))
else:
### We have monitors without any UI settable parameter.
monitors_list.append(self.available_monitors[str(monitor_name)]())
if len(monitors) < 1:
raise LaunchException("Can not launch operation without monitors selected !!!")
self.log.debug("%s: Initializing Coupling..." % str(self))
coupling_inst = self.available_couplings[str(coupling)](**coupling_parameters)
self.log.debug("Initializing Cortex...")
if self._is_surface_simulation(surface, surface_parameters):
cortex_entity = Cortex(use_storage=False).populate_cortex(surface, surface_parameters)
if cortex_entity.region_mapping_data.connectivity.number_of_regions != connectivity.number_of_regions:
raise LaunchException("Incompatible RegionMapping -- Connectivity !!")
if cortex_entity.region_mapping_data.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException("Incompatible RegionMapping -- Surface !!")
select_loc_conn = cortex_entity.local_connectivity
if select_loc_conn is not None and select_loc_conn.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException("Incompatible LocalConnectivity -- Surface !!")
else:
cortex_entity = None
self.log.debug("%s: Instantiating requested simulator..." % str(self))
if conduction_speed not in (0.0, None):
connectivity.speed = numpy.array([conduction_speed])
else:
raise LaunchException("conduction speed cannot be 0 or missing")
self.algorithm = Simulator(connectivity=connectivity, coupling=coupling_inst, surface=cortex_entity,
stimulus=stimulus, model=model_instance, integrator=integr,
monitors=monitors_list, initial_conditions=initial_conditions,
conduction_speed=conduction_speed)
self.simulation_length = simulation_length
self.log.debug("%s: Initializing storage..." % str(self))
try:
self.algorithm.preconfigure()
except ValueError as err:
raise LaunchException("Failed to configure simulator due to invalid Input Values. It could be because "
"of an incompatibility between different version of TVB code.", err)
def get_required_memory_size(self, **kwargs):
"""
Return the required memory to run this algorithm.
"""
return self.algorithm.memory_requirement()
def get_required_disk_size(self, **kwargs):
"""
Return the required disk size this algorithm estimates it will take. (in kB)
"""
return self.algorithm.storage_requirement(self.simulation_length) / 2 ** 10
def get_execution_time_approximation(self, **kwargs):
"""
Method should approximate based on input arguments, the time it will take for the operation
to finish (in seconds).
"""
# This is just a brute approx so cluster nodes won't kill operation before
# it's finished. This should be done with a higher grade of sensitivity
# Magic number connecting simulation length to simulation computation time
# This number should as big as possible, as long as it is still realistic, to
magic_number = 6.57e-06 # seconds
approx_number_of_nodes = 500
approx_nvar = 15
approx_modes = 15
simulation_length = int(float(kwargs['simulation_length']))
approx_integrator_dt = float(kwargs['integrator_parameters']['dt'])
if approx_integrator_dt == 0.0:
approx_integrator_dt = 1.0
if 'surface' in kwargs and kwargs['surface'] is not None and kwargs['surface'] != '':
approx_number_of_nodes *= approx_number_of_nodes
estimation = magic_number * approx_number_of_nodes * approx_nvar * approx_modes * simulation_length \
/ approx_integrator_dt
return max(int(estimation), 1)
def _try_find_mapping(self, mapping_class, connectivity_gid):
"""
Try to find a DataType instance of class "mapping_class", linked to the given Connectivity.
Entities in the current project will have priority.
:param mapping_class: DT class, with field "_connectivity" on it
:param connectivity_gid: GUID
:return: None or instance of "mapping_class"
"""
dts_list = dao.get_generic_entity(mapping_class, connectivity_gid, '_connectivity')
if len(dts_list) < 1:
return None
for dt in dts_list:
dt_operation = dao.get_operation_by_id(dt.fk_from_operation)
if dt_operation.fk_launched_in == self.current_project_id:
return dt
return dts_list[0]
def launch(self, model, model_parameters, integrator, integrator_parameters, connectivity,
monitors, monitors_parameters=None, surface=None, surface_parameters=None, stimulus=None,
coupling=None, coupling_parameters=None, initial_conditions=None,
conduction_speed=None, simulation_length=0, simulation_state=None):
"""
Called from the GUI to launch a simulation.
*: string class name of chosen model, etc...
*_parameters: dictionary of parameters for chosen model, etc...
connectivity: tvb.datatypes.connectivity.Connectivity object.
surface: tvb.datatypes.surfaces.CorticalSurface: or None.
stimulus: tvb.datatypes.patters.* object
"""
result_datatypes = dict()
start_time = self.algorithm.current_step * self.algorithm.integrator.dt
self.algorithm.configure(full_configure=False)
if simulation_state is not None:
simulation_state.fill_into(self.algorithm)
region_map = self._try_find_mapping(region_mapping.RegionMapping, connectivity.gid)
region_volume_map = self._try_find_mapping(region_mapping.RegionVolumeMapping, connectivity.gid)
for monitor in self.algorithm.monitors:
m_name = monitor.__class__.__name__
ts = monitor.create_time_series(self.storage_path, connectivity, surface, region_map, region_volume_map)
self.log.debug("Monitor %s created the TS %s" % (m_name, ts))
# Now check if the monitor will return results for each state variable, in which case store
# the labels for these state variables.
# todo move these into monitors as well
# and replace check if ts.user_tag_1 with something better (e.g. pre_ex & post)
state_variable_dimension_name = ts.labels_ordering[1]
if ts.user_tag_1:
ts.labels_dimensions[state_variable_dimension_name] = ts.user_tag_1.split(';')
elif m_name in self.HAVE_STATE_VARIABLES:
selected_vois = [self.algorithm.model.variables_of_interest[idx] for idx in monitor.voi]
ts.labels_dimensions[state_variable_dimension_name] = selected_vois
ts.start_time = start_time
result_datatypes[m_name] = ts
#### Create Simulator State entity and persist it in DB. H5 file will be empty now.
if not self._is_group_launch():
simulation_state = SimulationState(storage_path=self.storage_path)
self._capture_operation_results([simulation_state])
### Run simulation
self.log.debug("%s: Starting simulation..." % str(self))
for result in self.algorithm(simulation_length=simulation_length):
for j, monitor in enumerate(monitors):
if result[j] is not None:
result_datatypes[monitor].write_time_slice([result[j][0]])
result_datatypes[monitor].write_data_slice([result[j][1]])
self.log.debug("%s: Completed simulation, starting to store simulation state " % str(self))
### Populate H5 file for simulator state. This step could also be done while running sim, in background.
if not self._is_group_launch():
simulation_state.populate_from(self.algorithm)
self._capture_operation_results([simulation_state])
self.log.debug("%s: Simulation state persisted, returning results " % str(self))
final_results = []
for result in result_datatypes.values():
result.close_file()
final_results.append(result)
self.log.info("%s: Adapter simulation finished!!" % str(self))
return final_results
def _validate_model_parameters(self, model_instance, connectivity, surface):
"""
Checks if the size of the model parameters is set correctly.
"""
ui_configurable_params = model_instance.ui_configurable_parameters
for param in ui_configurable_params:
param_value = eval('model_instance.' + param)
if isinstance(param_value, numpy.ndarray):
if len(param_value) == 1 or connectivity is None:
continue
if surface is not None:
if (len(param_value) != surface.number_of_vertices
and len(param_value) != connectivity.number_of_regions):
msg = str(surface.number_of_vertices) + ' or ' + str(connectivity.number_of_regions)
msg = self._get_exception_message(param, msg, len(param_value))
self.log.error(msg)
raise LaunchException(msg)
elif len(param_value) != connectivity.number_of_regions:
msg = self._get_exception_message(param, connectivity.number_of_regions, len(param_value))
self.log.error(msg)
raise LaunchException(msg)
@staticmethod
def _get_exception_message(param_name, expected_size, actual_size):
"""
Creates the message that will be displayed to the user when the size of a model parameter is incorrect.
"""
msg = "The length of the parameter '" + param_name + "' is not correct."
msg += " It is expected to be an array of length " + str(expected_size) + "."
msg += " It is an array of length " + str(actual_size) + "."
return msg
@staticmethod
def _is_surface_simulation(surface, surface_parameters):
"""
Is this a surface simulation?
"""
return surface is not None and surface_parameters is not None
class SimulatorAdapter(ABCAsynchronous):
"""
Interface between the Simulator and the Framework.
"""
_ui_name = "Simulation Core"
algorithm = None
available_models = get_traited_subclasses(Model)
available_monitors = get_traited_subclasses(Monitor)
available_integrators = get_traited_subclasses(Integrator)
available_couplings = get_traited_subclasses(Coupling)
# This is a list with the monitors that actually return multi dimensions for the state variable dimension.
# We exclude from this for example EEG, MEG or Bold which return
HAVE_STATE_VARIABLES = ["GlobalAverage", "SpatialAverage", "Raw", "SubSample", "TemporalAverage"]
def __init__(self):
super(SimulatorAdapter, self).__init__()
self.log.debug("%s: Initialized..." % str(self))
def get_input_tree(self):
"""
Return a list of lists describing the interface to the simulator. This
is used by the GUI to generate the menus and fields necessary for
defining a simulation.
"""
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface[self.INTERFACE_ATTRIBUTES]
# We should add as hidden the Simulator State attribute.
result.append({self.KEY_NAME: 'simulation_state', self.KEY_TYPE: SimulationState,
self.KEY_LABEL: "Continuation of", self.KEY_REQUIRED: False, self.KEY_UI_HIDE: True})
return result
def get_output(self):
"""
:returns: list of classes for possible results of the Simulator.
"""
return [time_series.TimeSeries]
def configure(self, model, model_parameters, integrator, integrator_parameters, connectivity,
monitors, monitors_parameters=None, surface=None, surface_parameters=None, stimulus=None,
coupling=None, coupling_parameters=None, initial_conditions=None,
conduction_speed=None, simulation_length=0, simulation_state=None):
"""
Make preparations for the adapter launch.
"""
self.log.debug("available_couplings: %s..." % str(self.available_couplings))
self.log.debug("coupling: %s..." % str(coupling))
self.log.debug("coupling_parameters: %s..." % str(coupling_parameters))
self.log.debug("%s: Initializing Model..." % str(self))
noise_framework.build_noise(model_parameters)
model_instance = self.available_models[str(model)](**model_parameters)
self._validate_model_parameters(model_instance, connectivity, surface)
self.log.debug("%s: Initializing Integration scheme..." % str(self))
noise_framework.build_noise(integrator_parameters)
integr = self.available_integrators[integrator](**integrator_parameters)
self.log.debug("%s: Instantiating Monitors..." % str(self))
monitors_list = []
for monitor_name in monitors:
if (monitors_parameters is not None) and (str(monitor_name) in monitors_parameters):
current_monitor_parameters = monitors_parameters[str(monitor_name)]
HRFKernelEquation.build_equation_from_dict('hrf_kernel', current_monitor_parameters, True)
monitors_list.append(self.available_monitors[str(monitor_name)](**current_monitor_parameters))
else:
### We have monitors without any UI settable parameter.
monitors_list.append(self.available_monitors[str(monitor_name)]())
if len(monitors) < 1:
raise LaunchException("Can not launch operation without monitors selected !!!")
self.log.debug("%s: Initializing Coupling..." % str(self))
coupling_inst = self.available_couplings[str(coupling)](**coupling_parameters)
self.log.debug("Initializing Cortex...")
if self._is_surface_simulation(surface, surface_parameters):
cortex_entity = Cortex(use_storage=False).populate_cortex(surface, surface_parameters)
if cortex_entity.region_mapping_data.connectivity.number_of_regions != connectivity.number_of_regions:
raise LaunchException("Incompatible RegionMapping -- Connectivity !!")
if cortex_entity.region_mapping_data.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException("Incompatible RegionMapping -- Surface !!")
select_loc_conn = cortex_entity.local_connectivity
if select_loc_conn is not None and select_loc_conn.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException("Incompatible LocalConnectivity -- Surface !!")
else:
cortex_entity = None
self.log.debug("%s: Instantiating requested simulator..." % str(self))
if conduction_speed not in (0.0, None):
connectivity.speed = numpy.array([conduction_speed])
else:
raise LaunchException("conduction speed cannot be 0 or missing")
self.algorithm = Simulator(connectivity=connectivity, coupling=coupling_inst, surface=cortex_entity,
stimulus=stimulus, model=model_instance, integrator=integr,
monitors=monitors_list, initial_conditions=initial_conditions,
conduction_speed=conduction_speed)
self.simulation_length = simulation_length
self.log.debug("%s: Initializing storage..." % str(self))
try:
self.algorithm.preconfigure()
except ValueError, err:
raise LaunchException("Failed to configure simulator due to invalid Input Values. It could be because "
"of an incompatibility between different version of TVB code.", err)
class SimulatorAdapter(ABCAsynchronous):
"""
Interface between the Simulator and the Framework.
"""
_ui_name = "Simulation Core"
algorithm = None
available_models = get_traited_subclasses(Model)
available_monitors = get_traited_subclasses(Monitor)
available_integrators = get_traited_subclasses(Integrator)
available_couplings = get_traited_subclasses(Coupling)
# This is a list with the monitors that actually return multi dimensions for the state variable dimension.
# We exclude from this for example EEG, MEG or Bold which return
HAVE_STATE_VARIABLES = [
"GlobalAverage", "SpatialAverage", "Raw", "SubSample",
"TemporalAverage"
]
def __init__(self):
super(SimulatorAdapter, self).__init__()
self.log.debug("%s: Initialized..." % str(self))
def get_input_tree2(self):
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface_experimental
return result
def get_input_tree(self):
"""
Return a list of lists describing the interface to the simulator. This
is used by the GUI to generate the menus and fields necessary for
defining a simulation.
"""
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface[self.INTERFACE_ATTRIBUTES]
# We should add as hidden the Simulator State attribute.
result.append({
self.KEY_NAME: 'simulation_state',
self.KEY_TYPE: 'tvb.datatypes.simulation_state.SimulationState',
self.KEY_LABEL: "Continuation of",
self.KEY_REQUIRED: False,
self.KEY_UI_HIDE: True
})
return result
def get_output(self):
"""
:returns: list of classes for possible results of the Simulator.
"""
return [time_series.TimeSeries]
def configure(self,
model,
model_parameters,
integrator,
integrator_parameters,
connectivity,
monitors,
monitors_parameters=None,
surface=None,
surface_parameters=None,
stimulus=None,
coupling=None,
coupling_parameters=None,
initial_conditions=None,
conduction_speed=None,
simulation_length=0,
simulation_state=None):
"""
Make preparations for the adapter launch.
"""
self.log.debug("available_couplings: %s..." %
str(self.available_couplings))
self.log.debug("coupling: %s..." % str(coupling))
self.log.debug("coupling_parameters: %s..." % str(coupling_parameters))
self.log.debug("%s: Initializing Model..." % str(self))
noise_framework.build_noise(model_parameters)
model_instance = self.available_models[str(model)](**model_parameters)
self._validate_model_parameters(model_instance, connectivity, surface)
self.log.debug("%s: Initializing Integration scheme..." % str(self))
noise_framework.build_noise(integrator_parameters)
integr = self.available_integrators[integrator](
**integrator_parameters)
self.log.debug("%s: Instantiating Monitors..." % str(self))
monitors_list = []
for monitor_name in monitors:
if (monitors_parameters is not None) and (str(monitor_name)
in monitors_parameters):
current_monitor_parameters = monitors_parameters[str(
monitor_name)]
HRFKernelEquation.build_equation_from_dict(
'hrf_kernel', current_monitor_parameters, True)
monitors_list.append(
self.available_monitors[str(monitor_name)](
**current_monitor_parameters))
else:
### We have monitors without any UI settable parameter.
monitors_list.append(
self.available_monitors[str(monitor_name)]())
if len(monitors) < 1:
raise LaunchException(
"Can not launch operation without monitors selected !!!")
self.log.debug("%s: Initializing Coupling..." % str(self))
coupling_inst = self.available_couplings[str(coupling)](
**coupling_parameters)
self.log.debug("Initializing Cortex...")
if self._is_surface_simulation(surface, surface_parameters):
cortex_entity = Cortex(use_storage=False).populate_cortex(
surface, surface_parameters)
if cortex_entity.region_mapping_data.connectivity.number_of_regions != connectivity.number_of_regions:
raise LaunchException(
"Incompatible RegionMapping -- Connectivity !!")
if cortex_entity.region_mapping_data.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException(
"Incompatible RegionMapping -- Surface !!")
select_loc_conn = cortex_entity.local_connectivity
if select_loc_conn is not None and select_loc_conn.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException(
"Incompatible LocalConnectivity -- Surface !!")
else:
cortex_entity = None
self.log.debug("%s: Instantiating requested simulator..." % str(self))
if conduction_speed not in (0.0, None):
connectivity.speed = numpy.array([conduction_speed])
else:
raise LaunchException("conduction speed cannot be 0 or missing")
self.algorithm = Simulator(connectivity=connectivity,
coupling=coupling_inst,
surface=cortex_entity,
stimulus=stimulus,
model=model_instance,
integrator=integr,
monitors=monitors_list,
initial_conditions=initial_conditions,
conduction_speed=conduction_speed)
self.simulation_length = simulation_length
self.log.debug("%s: Initializing storage..." % str(self))
try:
self.algorithm.preconfigure()
except ValueError as err:
raise LaunchException(
"Failed to configure simulator due to invalid Input Values. It could be because "
"of an incompatibility between different version of TVB code.",
err)
def get_required_memory_size(self, **kwargs):
"""
Return the required memory to run this algorithm.
"""
return self.algorithm.memory_requirement()
def get_required_disk_size(self, **kwargs):
"""
Return the required disk size this algorithm estimates it will take. (in kB)
"""
return self.algorithm.storage_requirement(
self.simulation_length) / 2**10
def get_execution_time_approximation(self, **kwargs):
"""
Method should approximate based on input arguments, the time it will take for the operation
to finish (in seconds).
"""
# This is just a brute approx so cluster nodes won't kill operation before
# it's finished. This should be done with a higher grade of sensitivity
# Magic number connecting simulation length to simulation computation time
# This number should as big as possible, as long as it is still realistic, to
magic_number = 6.57e-06 # seconds
approx_number_of_nodes = 500
approx_nvar = 15
approx_modes = 15
simulation_length = int(float(kwargs['simulation_length']))
approx_integrator_dt = float(kwargs['integrator_parameters']['dt'])
if approx_integrator_dt == 0.0:
approx_integrator_dt = 1.0
if 'surface' in kwargs and kwargs[
'surface'] is not None and kwargs['surface'] != '':
approx_number_of_nodes *= approx_number_of_nodes
estimation = magic_number * approx_number_of_nodes * approx_nvar * approx_modes * simulation_length \
/ approx_integrator_dt
return max(int(estimation), 1)
def _try_find_mapping(self, mapping_class, connectivity_gid):
"""
Try to find a DataType instance of class "mapping_class", linked to the given Connectivity.
Entities in the current project will have priority.
:param mapping_class: DT class, with field "_connectivity" on it
:param connectivity_gid: GUID
:return: None or instance of "mapping_class"
"""
dts_list = dao.get_generic_entity(mapping_class, connectivity_gid,
'_connectivity')
if len(dts_list) < 1:
return None
for dt in dts_list:
dt_operation = dao.get_operation_by_id(dt.fk_from_operation)
if dt_operation.fk_launched_in == self.current_project_id:
return dt
return dts_list[0]
def launch(self,
model,
model_parameters,
integrator,
integrator_parameters,
connectivity,
monitors,
monitors_parameters=None,
surface=None,
surface_parameters=None,
stimulus=None,
coupling=None,
coupling_parameters=None,
initial_conditions=None,
conduction_speed=None,
simulation_length=0,
simulation_state=None):
"""
Called from the GUI to launch a simulation.
*: string class name of chosen model, etc...
*_parameters: dictionary of parameters for chosen model, etc...
connectivity: tvb.datatypes.connectivity.Connectivity object.
surface: tvb.datatypes.surfaces.CorticalSurface: or None.
stimulus: tvb.datatypes.patters.* object
"""
result_datatypes = dict()
start_time = self.algorithm.current_step * self.algorithm.integrator.dt
self.algorithm.configure(full_configure=False)
if simulation_state is not None:
simulation_state.fill_into(self.algorithm)
region_map = self._try_find_mapping(region_mapping.RegionMapping,
connectivity.gid)
region_volume_map = self._try_find_mapping(
region_mapping.RegionVolumeMapping, connectivity.gid)
for monitor in self.algorithm.monitors:
m_name = monitor.__class__.__name__
ts = monitor.create_time_series(self.storage_path, connectivity,
surface, region_map,
region_volume_map)
self.log.debug("Monitor %s created the TS %s" % (m_name, ts))
# Now check if the monitor will return results for each state variable, in which case store
# the labels for these state variables.
# todo move these into monitors as well
# and replace check if ts.user_tag_1 with something better (e.g. pre_ex & post)
state_variable_dimension_name = ts.labels_ordering[1]
if ts.user_tag_1:
ts.labels_dimensions[
state_variable_dimension_name] = ts.user_tag_1.split(';')
elif m_name in self.HAVE_STATE_VARIABLES:
selected_vois = [
self.algorithm.model.variables_of_interest[idx]
for idx in monitor.voi
]
ts.labels_dimensions[
state_variable_dimension_name] = selected_vois
ts.start_time = start_time
result_datatypes[m_name] = ts
#### Create Simulator State entity and persist it in DB. H5 file will be empty now.
if not self._is_group_launch():
simulation_state = SimulationState(storage_path=self.storage_path)
self._capture_operation_results([simulation_state])
### Run simulation
self.log.debug("%s: Starting simulation..." % str(self))
for result in self.algorithm(simulation_length=simulation_length):
for j, monitor in enumerate(monitors):
if result[j] is not None:
result_datatypes[monitor].write_time_slice([result[j][0]])
result_datatypes[monitor].write_data_slice([result[j][1]])
self.log.debug(
"%s: Completed simulation, starting to store simulation state " %
str(self))
### Populate H5 file for simulator state. This step could also be done while running sim, in background.
if not self._is_group_launch():
simulation_state.populate_from(self.algorithm)
self._capture_operation_results([simulation_state])
self.log.debug("%s: Simulation state persisted, returning results " %
str(self))
final_results = []
for result in result_datatypes.values():
result.close_file()
final_results.append(result)
self.log.info("%s: Adapter simulation finished!!" % str(self))
return final_results
def _validate_model_parameters(self, model_instance, connectivity,
surface):
"""
Checks if the size of the model parameters is set correctly.
"""
ui_configurable_params = model_instance.ui_configurable_parameters
for param in ui_configurable_params:
param_value = eval('model_instance.' + param)
if isinstance(param_value, numpy.ndarray):
if len(param_value) == 1 or connectivity is None:
continue
if surface is not None:
if (len(param_value) != surface.number_of_vertices
and len(param_value) !=
connectivity.number_of_regions):
msg = str(surface.number_of_vertices) + ' or ' + str(
connectivity.number_of_regions)
msg = self._get_exception_message(
param, msg, len(param_value))
self.log.error(msg)
raise LaunchException(msg)
elif len(param_value) != connectivity.number_of_regions:
msg = self._get_exception_message(
param, connectivity.number_of_regions,
len(param_value))
self.log.error(msg)
raise LaunchException(msg)
@staticmethod
def _get_exception_message(param_name, expected_size, actual_size):
"""
Creates the message that will be displayed to the user when the size of a model parameter is incorrect.
"""
msg = "The length of the parameter '" + param_name + "' is not correct."
msg += " It is expected to be an array of length " + str(
expected_size) + "."
msg += " It is an array of length " + str(actual_size) + "."
return msg
@staticmethod
def _is_surface_simulation(surface, surface_parameters):
"""
Is this a surface simulation?
"""
return surface is not None and surface_parameters is not None
class SimulatorAdapter(ABCAsynchronous):
"""
Interface between the Simulator and the Framework.
"""
_ui_name = "Simulation Core"
algorithm = None
available_models = get_traited_subclasses(Model)
available_monitors = get_traited_subclasses(Monitor)
available_integrators = get_traited_subclasses(Integrator)
available_couplings = get_traited_subclasses(Coupling)
# This is a list with the monitors that actually return multi dimensions for the state variable dimension.
# We exclude from this for example EEG, MEG or Bold which return
HAVE_STATE_VARIABLES = [
"GlobalAverage", "SpatialAverage", "Raw", "SubSample",
"TemporalAverage"
]
def __init__(self):
super(SimulatorAdapter, self).__init__()
self.log.debug("%s: Initialized..." % str(self))
def get_input_tree(self):
"""
Return a list of lists describing the interface to the simulator. This
is used by the GUI to generate the menus and fields necessary for
defining a simulation.
"""
sim = Simulator()
sim.trait.bound = self.INTERFACE_ATTRIBUTES_ONLY
result = sim.interface[self.INTERFACE_ATTRIBUTES]
# We should add as hidden the Simulator State attribute.
result.append({
self.KEY_NAME: 'simulation_state',
self.KEY_TYPE: SimulationState,
self.KEY_LABEL: "Continuation of",
self.KEY_REQUIRED: False,
self.KEY_UI_HIDE: True
})
return result
def get_output(self):
"""
:returns: list of classes for possible results of the Simulator.
"""
return [time_series.TimeSeries]
def configure(self,
model,
model_parameters,
integrator,
integrator_parameters,
connectivity,
monitors,
monitors_parameters=None,
surface=None,
surface_parameters=None,
stimulus=None,
coupling=None,
coupling_parameters=None,
initial_conditions=None,
conduction_speed=None,
simulation_length=0,
simulation_state=None):
"""
Make preparations for the adapter launch.
"""
self.log.debug("available_couplings: %s..." %
str(self.available_couplings))
self.log.debug("coupling: %s..." % str(coupling))
self.log.debug("coupling_parameters: %s..." % str(coupling_parameters))
self.log.debug("%s: Initializing Model..." % str(self))
noise_framework.build_noise(model_parameters)
model_instance = self.available_models[str(model)](**model_parameters)
self._validate_model_parameters(model_instance, connectivity, surface)
self.log.debug("%s: Initializing Integration scheme..." % str(self))
noise_framework.build_noise(integrator_parameters)
integr = self.available_integrators[integrator](
**integrator_parameters)
self.log.debug("%s: Instantiating Monitors..." % str(self))
monitors_list = []
for monitor_name in monitors:
if (monitors_parameters is not None) and (str(monitor_name)
in monitors_parameters):
current_monitor_parameters = monitors_parameters[str(
monitor_name)]
HRFKernelEquation.build_equation_from_dict(
'hrf_kernel', current_monitor_parameters, True)
monitors_list.append(
self.available_monitors[str(monitor_name)](
**current_monitor_parameters))
else:
### We have monitors without any UI settable parameter.
monitors_list.append(
self.available_monitors[str(monitor_name)]())
if len(monitors) < 1:
raise LaunchException(
"Can not launch operation without monitors selected !!!")
self.log.debug("%s: Initializing Coupling..." % str(self))
coupling_inst = self.available_couplings[str(coupling)](
**coupling_parameters)
self.log.debug("Initializing Cortex...")
if self._is_surface_simulation(surface, surface_parameters):
cortex_entity = Cortex(use_storage=False).populate_cortex(
surface, surface_parameters)
if cortex_entity.region_mapping_data.connectivity.number_of_regions != connectivity.number_of_regions:
raise LaunchException(
"Incompatible RegionMapping -- Connectivity !!")
if cortex_entity.region_mapping_data.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException(
"Incompatible RegionMapping -- Surface !!")
select_loc_conn = cortex_entity.local_connectivity
if select_loc_conn is not None and select_loc_conn.surface.number_of_vertices != surface.number_of_vertices:
raise LaunchException(
"Incompatible LocalConnectivity -- Surface !!")
else:
cortex_entity = None
self.log.debug("%s: Instantiating requested simulator..." % str(self))
if conduction_speed not in (0.0, None):
connectivity.speed = numpy.array([conduction_speed])
else:
raise LaunchException("conduction speed cannot be 0 or missing")
self.algorithm = Simulator(connectivity=connectivity,
coupling=coupling_inst,
surface=cortex_entity,
stimulus=stimulus,
model=model_instance,
integrator=integr,
monitors=monitors_list,
initial_conditions=initial_conditions,
conduction_speed=conduction_speed)
self.simulation_length = simulation_length
self.log.debug("%s: Initializing storage..." % str(self))
try:
self.algorithm.preconfigure()
except ValueError, err:
raise LaunchException(
"Failed to configure simulator due to invalid Input Values. It could be because "
"of an incompatibility between different version of TVB code.",
err)
