Closed j-wags closed 3 years ago
After some thought, I think we don't actually need this feature. Molecules that fail coverage report generation with one FF are like 99.99% going to fail with another, so as long as we run this once it doesn't really matter which one we use. Also, it's not clear what analysis could be done with multiple FFs' coverage reports. If we DO end up needing coverage reports for each different FF later, it would just take a few minutes to run separately for each.
Different versions of the force fields can differ slightly in their number of parameters, so in the final analysis we'll probably have a coverage report for EACH {molecule set, force field} combination. Right now the coverage-report step defaults to call the output file
3-coverage_report/coverage_report.json
, but it may be better to have it take as input a range of force fields, and then output a different JSON for each.