RLG: Runlog Generator

The Runlog Generator reads entire data stream for a single run and produces a RLG data structure that is written to the database. It should be run using:-

$processor_list inp upk cal rlg end

It must be fed all the sub-runs of a run.

This processor will run as part of the data production process on site; its unlikely to be useful elsewhere.

It creates the following RLG data structure:-

    MAST
        perm
        RLOG
            HIDx

where:-

MAST

is the top level bank and just acts as a hook for the banks it supports.

perm

represents selected permanent banks, such as RHDR and TRIG, that are copied in full. The data stream ordering is preserved - as new banks are encountered they are added to the end of the appropriate chain. The selection of banks that are copied like this is controlled by _MAST_MASK section of the

TRLG title bank

RLOG

This bank is divided into two sections:-

Hardwired

The first section is hardwired and contains such items as the date range and the maximum and minimum event numbers. Modifications to this section of the bank can only be made by updating the RLG code.

Collapsed n-tuple

The trailing section of the bank is build by NTP - the n-tuple builder, using the

RLNR 1000 title bank

The RLNR are n-tuple banks just like NTPR. RLNR 1000 holds a collapsed n-tuple (see section 6.9) which counts selected values from the input data structure, such as the number of times a particular trigger bit was set. The collapsed n-tuple is appended to the RLOG bank during termination and so this part of the RLOG bank can be updated by updating the RLNR 1000 bank.

HIDx

This is a chain of banks holding histograms and n-tuples in HBOOK format. In this form they are not amenable to analysis, but can be written to an HBOOK file ready for PAW, by RLO, the Runlog Output processor. All the histograms and n-tuples are generated using RLNR banks, so can be changed simply by modifying the current set of RLNR bank or by adding new ones.

As currently implemented, RLG outputs all the histograms and n-tuples that it finds in the HBOOK /PAWC/ common.

If RLG is run with other processors that create histograms, they will be stored in the RLG data structure!

The TRLG bank contains switches to control the writing of the RLG data structure to the database and to update DT2R and R2DT tables that are used to convert between date/time and run number. If the output is not written to the database then the histograms and n-tuples are written to an HBOOK file instead. By default the name of this file is rlg.hbk although this can be overridden using a FILE RLG 1 command.