Slowcontrol
In general, Ant runs event as a stream through the reconstruction, using calibration data which typically change for each file, and hands over nicely usable events to physics classes. See Architecture for details.
This page focuses on the insertion of "slow control data" within Ant, which changes more frequently than calibration data, but spans some thousands events. The data is processed and provided as some nicely usable entities to the physics classes, opposed to calibration information only used by the Reconstruction.
The main code parts reside in src/analysis/slowcontrol, and the PhysicsManager uses a SlowControlManager to handle the buffering of events.
The unpacker stage may add TSlowControl items with differing payload and types to a TEventData, found in the subfolder src/tree/. This is the representation to transfer this data to the slow control processing.
Processors look at reconstructed TEventData contents (mainly TSlowControl items) and turn them into physically meaningful data by knowing how to handle the Payload. Also, the give some feedback to the manager if enough events have been processed in order to represent a valid datum.
Those entitites use slow control processors to provide slow control information to the physics classes. They're typically Requested in the constructor of physics classes, and they have accessor methods for providing the well-formatted information.
The data acquisition in A2 saves scaler information, which are free running counters, every 10000 events. Their validity is backwards, as shown:
Since at the beginning of a run, the scalers might not be cleared, the block of events before the first scaler read is discarded automatically by Ant's design of AcquScaler slowcontrol processing.