With the main components in place, we can start running tuvx and returning values
Acceptance criteria
create from file/string, Run, the count getters, the three orderings.
- Tests build from a known-good config (reuse existing TUV-x test configs/data)
and check output shapes / orderings. Run actually executes here, so the
"good inputs only" rule matters most in this PR.
- if possible and it makes sense, format the data into a netcdf dataset
Ideas
- evaluate python and fortran interface implementations and see what fits best
With the main components in place, we can start running tuvx and returning values
Acceptance criteria
createfrom file/string,Run, the count getters, the three orderings.and check output shapes / orderings.
Runactually executes here, so the"good inputs only" rule matters most in this PR.
Ideas