Skip to content

Change of extracting needed signatures to import relevant files to make use of ChainDict

Catrin Bernius requested to merge bernius/athena:ATR20439_GenMenu into master

This MR presents a way to extract information which signature to import from the dictionary directly. This was something that @tbold had pointed out initially in ATR-20439 that there was unnecessary complexity in the menu code.

Some notes that can be also found on the jira but pointing some of them out again here:

  • The lists at the beginning of GenerateMenuMT.py with the various signatures might not be needed anymore perhaps or there is a different way to deal with the checks that the correct signature has been defined. In principle the dictionary doesn't rely on the information from which slice the chain is taken in the menu file.
  • The Bphysics dictionary had a key with the name subFolder instead of sigFolder, this was changed to include this. I didn't find an instance of the key being used anywhere in the code. Tag to @lyubushk and @abarton.
  • I'm not sure why there is a defaultSignature defined, and why this is the streaming one? Is this so that we always run at least a noalg chain? But what if there is none defined in the menu? Is this really needed? This had been always added as default to the menu but might have been a consequence of the situation described in the next bullet point.
  • With the previous way of handling things, the list of available signatures contained the "Cosmic" signature. This is not the case anymore. The reason being that there were three chains (HLT_noalg_SCTPEB_L1RD0_EMPTY, HLT_noalg_laser_TilePEB_L1CALREQ2, HLT_noalg_CIS_TilePEB_L1CALREQ1) added in the Cosmics slice of the menu I assume, but they're really streamers so are defined in the dictionary under Streaming. If this poses a problem, we'll need to find a fix for this still.

Tagging @dzanzi and @khoo as menu coordinators.

Merge request reports