Source code for tvb.adapters.visualizers.sensors

# -*- 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 to download
# TheVirtualBrain-Scientific Package (for simulators). See content of the
# documentation-folder for more details. See also http://www.thevirtualbrain.org
#
# (c) 2012-2024, 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 explained here:
# https://www.thevirtualbrain.org/tvb/zwei/neuroscience-publications
#
#

"""
.. moduleauthor:: Lia Domide <lia.domide@codemart.ro>
.. moduleauthor:: Mihai Andrei <mihai.andrei@codemart.ro>
"""

import json

from tvb.adapters.datatypes.db.sensors import SensorsIndex
from tvb.adapters.visualizers.surface_view import ensure_shell_surface, SurfaceURLGenerator
from tvb.basic.logger.builder import get_logger
from tvb.core.adapters.abcadapter import ABCAdapterForm
from tvb.core.adapters.abcdisplayer import ABCDisplayer, URLGenerator
from tvb.core.adapters.exceptions import LaunchException
from tvb.core.neocom import h5
from tvb.core.neotraits.forms import TraitDataTypeSelectField
from tvb.core.neotraits.view_model import ViewModel, DataTypeGidAttr
from tvb.datatypes.sensors import SensorsInternal, SensorsEEG, SensorsMEG, Sensors
from tvb.datatypes.surfaces import Surface, SurfaceTypesEnum

LOG = get_logger(__name__)


[docs] def prepare_sensors_as_measure_points_params(sensors): """ Returns urls from where to fetch the measure points and their labels """ sensor_locations = URLGenerator.build_h5_url(sensors.gid, 'get_locations') sensor_no = sensors.number_of_sensors sensor_labels = URLGenerator.build_h5_url(sensors.gid, 'get_labels') return {'urlMeasurePoints': sensor_locations, 'urlMeasurePointsLabels': sensor_labels, 'noOfMeasurePoints': sensor_no, 'minMeasure': 0, 'maxMeasure': sensor_no, 'urlMeasure': ''}
[docs] def prepare_mapped_sensors_as_measure_points_params(sensors, eeg_cap=None, adapter_id=None): """ Compute sensors positions by mapping them to the ``eeg_cap`` surface If ``eeg_cap`` is not specified the mapping will use a default EEGCal DataType in current project. If no default EEGCap is found, return sensors as they are (not projected) :returns: dictionary to be used in Viewers for rendering measure_points :rtype: dict """ if eeg_cap: sensor_locations = URLGenerator.build_url(adapter_id, 'sensors_to_surface', sensors.gid, parameter='surface_to_map_gid=' + eeg_cap.gid) sensor_no = sensors.number_of_sensors sensor_labels = URLGenerator.build_h5_url(sensors.gid, 'get_labels') return {'urlMeasurePoints': sensor_locations, 'urlMeasurePointsLabels': sensor_labels, 'noOfMeasurePoints': sensor_no, 'minMeasure': 0, 'maxMeasure': sensor_no, 'urlMeasure': ''} return prepare_sensors_as_measure_points_params(sensors)
[docs] def function_sensors_to_surface(sensors_gid, surface_to_map_gid): """ Map EEG sensors onto the head surface (skin-air). EEG sensor locations are typically only given on a unit sphere, that is, they are effectively only identified by their orientation with respect to a coordinate system. This method is used to map these unit vector sensor "locations" to a specific location on the surface of the skin. Assumes coordinate systems are aligned, i.e. common x,y,z and origin. """ sensors_dt = h5.load_from_gid(sensors_gid) surface_dt = h5.load_from_gid(surface_to_map_gid) return sensors_dt.sensors_to_surface(surface_dt).tolist()
[docs] class SensorsViewerModel(ViewModel): sensors = DataTypeGidAttr( linked_datatype=Sensors, label='Sensors', doc='Internals sensors to view' ) projection_surface = DataTypeGidAttr( linked_datatype=Surface, required=False, label='Projection Surface', doc='A surface on which to project the results. When missing, ' 'the first EEGCap is taken. This parameter is ignored when ' 'InternalSensors are inspected' ) shell_surface = DataTypeGidAttr( linked_datatype=Surface, required=False, label='Shell Surface', doc='Wrapping surface over the internal sensors, to be displayed ' 'semi-transparently, for visual purposes only.' )
[docs] class SensorsViewerForm(ABCAdapterForm): def __init__(self): super(SensorsViewerForm, self).__init__() self.sensors = TraitDataTypeSelectField(SensorsViewerModel.sensors, name='sensors', conditions=self.get_filters()) self.projection_surface = TraitDataTypeSelectField(SensorsViewerModel.projection_surface, name='projection_surface') self.shell_surface = TraitDataTypeSelectField(SensorsViewerModel.shell_surface, name='shell_surface')
[docs] @staticmethod def get_view_model(): return SensorsViewerModel
[docs] @staticmethod def get_required_datatype(): return SensorsIndex
[docs] @staticmethod def get_input_name(): return 'sensors'
[docs] @staticmethod def get_filters(): return None
[docs] class SensorsViewer(ABCDisplayer): """ Sensor visualizer - for visual inspecting of TVB Sensors DataTypes. """ _ui_name = "Sensor Visualizer" _ui_subsection = "sensors"
[docs] def get_form_class(self): return SensorsViewerForm
[docs] def launch(self, view_model): # type: (SensorsViewerModel) -> dict """ Prepare visualizer parameters. We support viewing all sensor types through a single viewer, so that a user doesn't need to go back to the data-page, for loading a different type of sensor. """ sensors_index = self.load_entity_by_gid(view_model.sensors) shell_surface_index = None projection_surface_index = None if view_model.shell_surface: shell_surface_index = self.load_entity_by_gid(view_model.shell_surface) if view_model.projection_surface: projection_surface_index = self.load_entity_by_gid(view_model.projection_surface) if sensors_index.sensors_type == SensorsInternal.sensors_type.default: return self._params_internal_sensors(sensors_index, shell_surface_index) if sensors_index.sensors_type == SensorsEEG.sensors_type.default: return self._params_eeg_sensors(sensors_index, projection_surface_index, shell_surface_index) if sensors_index.sensors_type == SensorsMEG.sensors_type.default: return self._params_meg_sensors(sensors_index, projection_surface_index, shell_surface_index) raise LaunchException("Unknown sensors type!")
def _params_internal_sensors(self, internal_sensors, shell_surface=None): params = prepare_sensors_as_measure_points_params(internal_sensors) shell_surface = ensure_shell_surface(self.current_project_id, shell_surface, SurfaceTypesEnum.CORTICAL_SURFACE.value) params['shellObject'] = self.prepare_shell_surface_params(shell_surface, SurfaceURLGenerator) return self.build_display_result('sensors/sensors_internal', params, pages={"controlPage": "sensors/sensors_controls"}) def _params_eeg_sensors(self, eeg_sensors, eeg_cap=None, shell_surface=None): if eeg_cap is None: eeg_cap = ensure_shell_surface(self.current_project_id, eeg_cap, SurfaceTypesEnum.EEG_CAP_SURFACE.value) params = prepare_mapped_sensors_as_measure_points_params(eeg_sensors, eeg_cap, self.stored_adapter.id) shell_surface = ensure_shell_surface(self.current_project_id, shell_surface) params.update({ 'shellObject': self.prepare_shell_surface_params(shell_surface, SurfaceURLGenerator), 'urlVertices': '', 'urlTriangles': '', 'urlLines': '[]', 'urlNormals': '' }) if eeg_cap is not None: with h5.h5_file_for_gid(eeg_cap.gid) as eeg_cap_h5: params.update(self._compute_surface_params(eeg_cap_h5)) return self.build_display_result("sensors/sensors_eeg", params, pages={"controlPage": "sensors/sensors_controls"}) def _params_meg_sensors(self, meg_sensors, projection_surface=None, shell_surface=None): params = prepare_sensors_as_measure_points_params(meg_sensors) shell_surface = ensure_shell_surface(self.current_project_id, shell_surface) params.update({ 'shellObject': self.prepare_shell_surface_params(shell_surface, SurfaceURLGenerator), 'urlVertices': '', 'urlTriangles': '', 'urlLines': '[]', 'urlNormals': '', 'boundaryURL': '', 'urlRegionMap': ''}) if projection_surface is not None: with h5.h5_file_for_gid(projection_surface.gid) as projection_surface_h5: params.update(self._compute_surface_params(projection_surface_h5)) return self.build_display_result("sensors/sensors_eeg", params, pages={"controlPage": "sensors/sensors_controls"}) @staticmethod def _compute_surface_params(surface_h5): rendering_urls = [json.dumps(url) for url in SurfaceURLGenerator.get_urls_for_rendering(surface_h5)] url_vertices, url_normals, url_lines, url_triangles, _ = rendering_urls return {'urlVertices': url_vertices, 'urlTriangles': url_triangles, 'urlLines': url_lines, 'urlNormals': url_normals}
[docs] def get_required_memory_size(self): return -1
[docs] def sensors_to_surface(self, sensors_gid, surface_to_map_gid): # Method needs to be defined on the adapter, to be called from JS return function_sensors_to_surface(sensors_gid, surface_to_map_gid)