filter_windows

Module: missions.nodes.data_selection.filter_windows

Filtering of windows based on the label

Inheritance diagram for pySPACE.missions.nodes.data_selection.filter_windows:

Class Summary

FilterGenerator(generator[, caching, ...])	Generator object that performs the actual filtering of the windows.
FilterWindowsNode([trigger_event, ...])	Filter out the windows depending on their label.

Classes

FilterGenerator

class pySPACE.missions.nodes.data_selection.filter_windows.FilterGenerator(generator, caching=False, trigger_event=None, activation_label=None, positive_event=None, negative_event=None, deactivation_label=None)

Bases: object

Generator object that performs the actual filtering of the windows.

Class Components Summary

_fetch_from_generator()	Fetches one fresh value from the generator, store it in the cache and yield it
filter_windows(value)	Method that is called for the filtering process.
fresh()	Return one generator that yields the same values like the internal one that was passed to __init__.

__init__(generator, caching=False, trigger_event=None, activation_label=None, positive_event=None, negative_event=None, deactivation_label=None)

Stores the generator and creates an empty cache

Note

Since the output of the generator is ordered, the cache is an ordered sequence of variable length like a list

_fetch_from_generator()

Fetches one fresh value from the generator, store it in the cache and yield it

fresh()

Return one generator that yields the same values like the internal one that was passed to __init__.

Note

It does not recompute values that have already been requested before but just uses these from the internal cache.

Note

Calling fresh invalidates all existing generators that have been created before using this method, i.e. there can only be one generator at a time

filter_windows(value)

Method that is called for the filtering process.

__weakref__

list of weak references to the object (if defined)

FilterWindowsNode

class pySPACE.missions.nodes.data_selection.filter_windows.FilterWindowsNode(trigger_event=None, activation_label=None, deactivation_label=None, positive_event=None, negative_event=None, use_test_data=True, **kwargs)

Bases: pySPACE.missions.nodes.base_node.BaseNode

Filter out the windows depending on their label.

The node inspects the label of the window and decides, based on the specified criteria and past events, if the window should be forwarded or ignored.

It is possible to use a simple switch on / switch off mechanism or more complex state-machine like filtering with triggers.

Parameters

activation_label:  

Start streaming windows after a window with this label.

(optional, default: None)

deactivation_label:  

Stop streaming windows after a window with this label.

(optional, default: None)

positive_event:

Positive event for the state machine.

(optional, default: None)

negative_event:

Negative event for the state machine.

(optional, default: None)

Exemplary Call

-
    node : FilterWindows
    parameters :
        deactivation_label : "Stop"
        activation_label : "Start"

Authors:

Hendrik Woehrle (hendrik.woehrle@dfki.de)

Created:

2013/04/04

POSSIBLE NODE NAMES:  

• FilterWindows
• FilterWindowsNode
• Filter_Windows

POSSIBLE INPUT TYPES:  

• PredictionVector
• FeatureVector
• TimeSeries

Class Components Summary

input_types
request_data_for_testing()	Returns the data for testing of subsequent nodes
request_data_for_training(use_test_data)	Returns the filtered data for training of subsequent nodes

__init__(trigger_event=None, activation_label=None, deactivation_label=None, positive_event=None, negative_event=None, use_test_data=True, **kwargs)

request_data_for_training(use_test_data)

Returns the filtered data for training of subsequent nodes

request_data_for_testing()

Returns the data for testing of subsequent nodes

input_types = ['PredictionVector', 'FeatureVector', 'TimeSeries']