live Package¶
live Package¶
Modules for the online usage of pySPACE
Contains the modules that are actually needed if pySPACE is used in any online application.
It wires all required parts together and configures them according to given configuration and parameterization files.
The pyspace live mode of operation is structured in several different epochs.
- In the optional prewindowing phase, pyspace live connects to a server that provides the data, performs preprocessing operations and stores the preprocessed data.
- The preprocessing operations have to be provided in a flow, and they should not be trainable. The trainable modules can only be used in the train phase.
- In the training phase, pyspace live loads stored data and uses it to train the trainable nodes of a flow.
- If the origin of the data is the prewindowing phase, this phase is called prewindowed train.
- In the optional adaptation phase, the threshold of the classifier is adapted to the actual task
- In the prediction phase, the trained flows are used to predict the class of new data, this
An introductory unit test is found in unit test directory.
Modules¶
adaptation |
Script for running threshold adaptation |
eeg_stream_manager |
Script for managing of eeg data streams |
online_utilities |
Utility functions for pyspace live, like special logging functionality |
prediction |
Script to run the actual online classification of data |
recorder |
|
trainer |
The module that trains pyspace flows. |
Subpackages Summary¶
communication |
Handle communication to other experiment components |