Currently in openff-benchmark optimize execute, execution occurs per spec in the specified season, iterating through all selected molecules. This is normally just fine, but if a user wants to re-run only a single spec in a season, the current code will execute all specs in that season for each molecule.
We can avoid this if we iterate at the top level through the molecules, and then for each molecule iterate through the seasons. If results are already present for that molecule and season, execution can skip.
Currently in
openff-benchmark optimize execute
, execution occurs per spec in the specified season, iterating through all selected molecules. This is normally just fine, but if a user wants to re-run only a single spec in a season, the current code will execute all specs in that season for each molecule.We can avoid this if we iterate at the top level through the molecules, and then for each molecule iterate through the seasons. If results are already present for that molecule and season, execution can skip.