In some circumstances, matching across into another simulation could result in
too many results being generated (and an exception being thrown), because at
the final stage the maximum weight criterion was not being applied correctly.
The resolution involves fixing an argmax to be grouped only by source halo,
not the most recent hop.
Added also a useful printing routine for debugging issues with multisource/multihop
calculations (disabled by default, of course!)
apontzen
commented
2 years ago
In some circumstances, matching across into another simulation could result in too many results being generated (and an exception being thrown), because at the final stage the maximum weight criterion was not being applied correctly.
The resolution involves fixing an argmax to be grouped only by source halo, not the most recent hop.
Added also a useful printing routine for debugging issues with multisource/multihop calculations (disabled by default, of course!)