Source/Sink

Sources and Sinks are the inputs/outputs of any pipelines. ORK provides a few for you but you can extend them by providing your own. For both, you now know the drill: you need a Python wrapper that wraps your Sourc/Sink cell. you just need to implement a Python object inheriting from some base classes.

For a Source: object_recognition_core.io.source.SourceBase:

class object_recognition_core.io.source.OpenNI(*args, **kwargs)

A source for any ORK pipeline, interface with a Kinect/ASUS Xtion pro using the openni driver

And for a Sink: py:class:object_recognition_core.io.source.SinkBase:

class object_recognition_core.io.sink.SinkBase

This is a base class for a sink: you don’t need to have your sink cell inherit from that class but if you do, it will make sure that its inputs/outputs fit the ORK standard (which is good if you want to interact with the official ORK pipelines). You need to call the BlackBox constructor in your __init__ first and then this function. Typically, yout __init__ is

>>> class Foo(ecto.BlackBox, SinkBase):
>>>     def __init__(self, *args, **kwargs):
>>>         ecto.BlackBox.__init__(self, *args, **kwargs)
>>>         SinkBase.__init__(self)