ru.primetalk.synapse.core.runtime.SignalProcessingApi0
Tracking processing collects lost and final traces during signal processing.
Tracking processing collects lost and final traces during signal processing. (Lost traces are those that haven't produced any result.) Be careful! Collecting traces in a system with intensive processing can quickly lead to memory overflow. This processing is intended for debug purposes.
The type that is used to represent a single value related to signal
TrellisBuilder is a way for a final user to watch the process of trellis building.
TrellisBuilder is a way for a final user to watch the process of trellis building. One may override #newTotalTrellisBuilder and put own implementation of TotalTrellisBuilder.
Constructs trellis.
Constructs trellis. After running a proc over the trace signal, we construct the next step of the trellis. An instance of this class is constructed anew for every trellis step.
Generates trellis until there are some data on nonStop contacts.
Generates trellis until there are some data on nonStop contacts. Can also process signals from child subsystems (not constrained only to input contacts). Processes one signal at a time. Has the same interface as RuntimeComponentMultiState.
It works as follows. Constructs a lazy evaluated Stream of TrellisElement s (method from
).
Then searches within the stream for a first element that contains only signals at stop contacts.
A component that does single step along the trellis.
Converts a runtime system to Trellis producer.
A generic trait for different signal processing methods. There are two descendants - a tracking one and a simple one. The tracking signal processor keeps track of what signals were produced during processing. The simple one do not save traces. So it requires less memory for processing.
Keeping traces can greatly help with debugging. One may see which signals and processors have lead to the result. It may be worth to save trellis to dot file.
Having a common ancestor for both types of processing is advantageous because there will be no code duplication among processors.