Configure deserialisation and decoding of the available HLT results in trigger reco

@tamartin, @fwinkl, I took a stab at:

• Determining the DS type (if applicable) from the input file stream name, and in the process tried to move us towards always using the functions in DataScoutingInfo for retrieval so that there are no risks of modifying the dict (needs follow-up based on LXR hits)
• Configuring deserialisation of all available HLT results at least when reading physics_Main BS. This is easy enough to configure, but we have to be careful about a couple of things:
  • Multiple HLT results cause warnings from attempting to write the same containers (in particular TLA jets)
  • We need to ensure the HLT results are all present in the file. Actually I just need to fix this I think:
    • Check the list of processing tags in the metadata to determine which streams were active.

It could be that I misunderstood the idea, but if not we may need to add an option to the TrigEDMDeserialiser such that it can gracefully fail on duplicated payloads and set this for the secondary results.

This also changes the TLA decoding to use RecoSteering in order to insert the GapFiller and address ATR-28247.