ATR-25328, ATR-25329: Correct spare threshold values to fit in L1Topo bits and avoid overflow of isolation cuts

Addresses the following.

• ATR-25328: Corrects Phase-I L1Talo SPARE menu thresholds, dividing all these by 10 to correct a typo and make sure they correspond to the max value in the available L1Topo bits encoding these values.
• ATR-25329: Removes use of "HadLoose/HadMedium/HadTight" working points for eTAU thresholds. Specifically, eTAU40HM was the only threshold cutting on rHad, and used "HadMedium" to specify the variable cut, which when translated to an enum value exceeded the available bits in L1Topo encoding the WP. As the rCore and rHad cuts are specified separately for each threshold, we can instead simply define corresponding cuts for both variables in "Loose/Medium/Tight", while maintaining the ability to cut on them independently.

The latter fix appears to also address the fact that eFEX simulation was itself unaware of the "Had" working points, which seems to have led to odd results: many events passing L1 but then failing to seed any HLT reconstruction. Could be that the "HadMedium" encoding interacted badly with the threshold pattern tool. Performance seems more reasonable with specifying the rHad cut via "Medium" (see JIRA).