TheVirtualBrain:

TheDocumentationwebsite.

Source code for tvb.adapters.simulator.simulator_adapter

# -*- coding: utf-8 -*-
#
#
# TheVirtualBrain-Framework Package. This package holds all Data Management, and 
# Web-UI helpful to run brain-simulations. To use it, you also need do download
# TheVirtualBrain-Scientific Package (for simulators). See content of the
# documentation-folder for more details. See also http://www.thevirtualbrain.org
#
# (c) 2012-2022, Baycrest Centre for Geriatric Care ("Baycrest") and others
#
# This program is free software: you can redistribute it and/or modify it under the
# terms of the GNU General Public License as published by the Free Software Foundation,
# either version 3 of the License, or (at your option) any later version.
# This program 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 General Public License for more details.
# You should have received a copy of the GNU General Public License along with this
# program.  If not, see <http://www.gnu.org/licenses/>.
#
#
#   CITATION:
# When using The Virtual Brain for scientific publications, please cite it as follows:
#
#   Paula Sanz Leon, Stuart A. Knock, M. Marmaduke Woodman, Lia Domide,
#   Jochen Mersmann, Anthony R. McIntosh, Viktor Jirsa (2013)
#       The Virtual Brain: a simulator of primate brain network dynamics.
#   Frontiers in Neuroinformatics (7:10. doi: 10.3389/fninf.2013.00010)
#
#

"""
Adapter that uses the traits module to generate interfaces to the Simulator.
Few supplementary steps are done here:

   * from submitted Monitor/Model... names, build transient entities
   * after UI parameters submit, compose transient Cortex entity to be passed to the Simulator.

.. moduleauthor:: Paula Popa <paula.popa@codemart.ro>
.. moduleauthor:: Lia Domide <lia.domide@codemart.ro>
.. moduleauthor:: Stuart A. Knock <Stuart@tvb.invalid>

"""

import json

from tvb.adapters.datatypes.db.connectivity import ConnectivityIndex
from tvb.adapters.datatypes.db.region_mapping import RegionMappingIndex, RegionVolumeMappingIndex
from tvb.adapters.datatypes.db.simulation_history import SimulationHistoryIndex
from tvb.adapters.datatypes.db.time_series import TimeSeriesIndex
from tvb.adapters.simulator.coupling_forms import get_ui_name_to_coupling_dict
from tvb.adapters.simulator.model_forms import get_model_to_form_dict
from tvb.adapters.simulator.monitor_forms import get_monitor_to_form_dict
from tvb.adapters.simulator.simulator_fragments import *
from tvb.basic.neotraits.api import Attr
from tvb.core.adapters.abcadapter import ABCAdapterForm, ABCAdapter
from tvb.core.adapters.exceptions import LaunchException, InvalidParameterException
from tvb.core.entities.file.simulator.simulation_history_h5 import SimulationHistory
from tvb.core.entities.file.simulator.view_model import SimulatorAdapterModel
from tvb.core.entities.storage import dao
from tvb.core.neocom import h5
from tvb.core.neotraits.forms import FloatField, SelectField
from tvb.simulator.coupling import Coupling
from tvb.simulator.simulator import Simulator


[docs]class SimulatorAdapterForm(ABCAdapterForm): def __init__(self): super(SimulatorAdapterForm, self).__init__() self.coupling_choices = get_ui_name_to_coupling_dict() default_coupling = list(self.coupling_choices.values())[0] self.connectivity = TraitDataTypeSelectField(SimulatorAdapterModel.connectivity, name=self.get_input_name(), conditions=self.get_filters()) self.coupling = SelectField( Attr(Coupling, default=default_coupling, label="Coupling", doc=Simulator.coupling.doc), name='coupling', choices=self.coupling_choices) self.conduction_speed = FloatField(Simulator.conduction_speed) self.ordered_fields = (self.connectivity, self.conduction_speed, self.coupling) self.range_params = [Simulator.connectivity, Simulator.conduction_speed]
[docs] def fill_from_trait(self, trait): # type: (Simulator) -> None if hasattr(trait, 'connectivity'): self.connectivity.data = trait.connectivity.hex self.coupling.data = trait.coupling.__class__ self.conduction_speed.data = trait.conduction_speed
[docs] def fill_trait(self, datatype): datatype.connectivity = self.connectivity.value datatype.conduction_speed = self.conduction_speed.value coupling = self.coupling.value if type(datatype.coupling) != coupling: datatype.coupling = coupling()
@staticmethod
[docs] def get_view_model(): return SimulatorAdapterModel
@staticmethod
[docs] def get_input_name(): return 'connectivity'
@staticmethod
[docs] def get_filters(): return None
@staticmethod
[docs] def get_required_datatype(): return ConnectivityIndex
def __str__(self): pass
[docs]class SimulatorAdapter(ABCAdapter): """ Interface between the Simulator and the Framework. """ _ui_name = "Simulation Core" algorithm = None branch_simulation_state_gid = None # 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))
[docs] def get_form_class(self): return SimulatorAdapterForm
[docs] def get_adapter_fragments(self, view_model): # type (SimulatorAdapterModel) -> dict forms = {None: [SimulatorStimulusFragment, SimulatorModelFragment, SimulatorIntegratorFragment, SimulatorMonitorFragment, SimulatorFinalFragment], "surface": [SimulatorSurfaceFragment, SimulatorRMFragment]} current_model_class = type(view_model.model) all_model_forms = get_model_to_form_dict() forms["model"] = [all_model_forms.get(current_model_class)] all_monitor_forms = get_monitor_to_form_dict() selected_monitor_forms = [] for monitor in view_model.monitors: current_monitor_class = type(monitor) selected_monitor_forms.append(all_monitor_forms.get(current_monitor_class)) forms["monitors"] = selected_monitor_forms # Not sure if where we should in fact include the entire tree, or it will become too tedious. # For now I think it is ok if we rename this section "Summary" and filter what is shown return forms
[docs] def get_output(self): """ :returns: list of classes for possible results of the Simulator. """ return [TimeSeriesIndex, SimulationHistoryIndex]
[docs] def configure(self, view_model): # type: (SimulatorAdapterModel) -> None """ Make preparations for the adapter launch. """ self.log.debug("%s: Configuring simulator adapter..." % str(self)) self.algorithm = self.view_model_to_has_traits(view_model) self.branch_simulation_state_gid = view_model.history_gid 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)
[docs] def get_required_memory_size(self, view_model): # type: (SimulatorAdapterModel) -> int """ Return the required memory to run this algorithm. """ return self.algorithm.memory_requirement()
[docs] def get_required_disk_size(self, view_model): # type: (SimulatorAdapterModel) -> int """ Return the required disk size this algorithm estimates it will take. (in kB) """ return self.algorithm.storage_requirement() / 2 ** 10
[docs] def get_execution_time_approximation(self, view_model): # type: (SimulatorAdapterModel) -> int """ 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 approx_integrator_dt = self.algorithm.integrator.dt if approx_integrator_dt == 0.0: approx_integrator_dt = 1.0 if self.algorithm.is_surface_simulation: approx_number_of_nodes *= approx_number_of_nodes estimation = (magic_number * approx_number_of_nodes * approx_nvar * approx_modes * self.algorithm.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, 'fk_connectivity_gid') 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 _try_load_region_mapping(self): region_map = None region_volume_map = None region_map_index = self._try_find_mapping(RegionMappingIndex, self.algorithm.connectivity.gid.hex) region_volume_map_index = self._try_find_mapping(RegionVolumeMappingIndex, self.algorithm.connectivity.gid.hex) if region_map_index: region_map = h5.load_from_index(region_map_index) if region_volume_map_index: region_volume_map = h5.load_from_index(region_volume_map_index) return region_map, region_volume_map
[docs] def launch(self, view_model): # type: (SimulatorAdapterModel) -> [TimeSeriesIndex, SimulationHistoryIndex] """ 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_h5 = dict() result_indexes = dict() start_time = self.algorithm.current_step * self.algorithm.integrator.dt self.algorithm.configure(full_configure=False) if self.branch_simulation_state_gid is not None: history = self.load_traited_by_gid(self.branch_simulation_state_gid) assert isinstance(history, SimulationHistory) history.fill_into(self.algorithm) region_map, region_volume_map = self._try_load_region_mapping() for monitor in self.algorithm.monitors: if monitor.period > view_model.simulation_length: raise InvalidParameterException("Sampling period for monitors can not be bigger " "than the simulation length!") m_name = type(monitor).__name__ ts = monitor.create_time_series(self.algorithm.connectivity, self.algorithm.surface, region_map, region_volume_map) self.log.debug("Monitor created the TS") ts.start_time = start_time ts_index_class = h5.REGISTRY.get_index_for_datatype(type(ts)) ts_index = ts_index_class() ts_index.fill_from_has_traits(ts) ts_index.data_ndim = 4 ts_index.state = 'INTERMEDIATE' state_variable_dimension_name = ts.labels_ordering[1] if 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_index.labels_dimensions = json.dumps(ts.labels_dimensions) ts_h5_class = h5.REGISTRY.get_h5file_for_datatype(type(ts)) ts_h5_path = h5.path_for(self._get_output_path(), ts_h5_class, ts.gid) self.log.info("Generating Timeseries at: {}".format(ts_h5_path)) ts_h5 = ts_h5_class(ts_h5_path) ts_h5.store(ts, scalars_only=True, store_references=False) ts_h5.sample_rate.store(ts.sample_rate) ts_h5.nr_dimensions.store(ts_index.data_ndim) # Storing GA also here redundant, except for HPC ts_h5.store_generic_attributes(self.generic_attributes) ts_h5.store_references(ts) result_indexes[m_name] = ts_index result_h5[m_name] = ts_h5 # Run simulation self.log.debug("Starting simulation...") for result in self.algorithm(simulation_length=self.algorithm.simulation_length): for j, monitor in enumerate(self.algorithm.monitors): if result[j] is not None: m_name = type(monitor).__name__ ts_h5 = result_h5[m_name] ts_h5.write_time_slice([result[j][0]]) ts_h5.write_data_slice([result[j][1]]) self.log.debug("Completed simulation, starting to store simulation state ") # Now store simulator history, at the simulation end results = [] if not self._is_group_launch(): simulation_history = SimulationHistory() simulation_history.populate_from(self.algorithm) self.generic_attributes.visible = False history_index = h5.store_complete(simulation_history, self._get_output_path(), self.generic_attributes) self.generic_attributes.visible = True history_index.fixed_generic_attributes = True results.append(history_index) self.log.debug("Simulation state persisted, returning results ") for monitor in self.algorithm.monitors: m_name = type(monitor).__name__ ts_shape = result_h5[m_name].read_data_shape() result_indexes[m_name].fill_shape(ts_shape) result_h5[m_name].close() self.log.debug("%s: Adapter simulation finished!!" % str(self)) results.extend(result_indexes.values()) return results