Problem Statement
For the TGLF stuff (as an example), they have an output file (eg CSV, parquet, hdf5, etc) where each row is a different simulation, and each column represents the final value of some parameter. To run that through nosim, we need to upload those as separate runs, with the values as metadata. We cannot do that through the normal python API because itll be too slow (ie couldn't loop through each row in the file, create a Run and POST it individually) because there are tens/hundreds of thousands of rows
Solution
We need some way to quickly do basically simvue run create --name=my_tglf_run --folder=/tglf --csv-file=results.csv
and then itll create a run for each row in the file, with name eg my_tglf_run_1, my_tglf_run_2 etc, with metadata equal to the data in that row of the file.
Problem Statement
For the TGLF stuff (as an example), they have an output file (eg CSV, parquet, hdf5, etc) where each row is a different simulation, and each column represents the final value of some parameter. To run that through nosim, we need to upload those as separate runs, with the values as metadata. We cannot do that through the normal python API because itll be too slow (ie couldn't loop through each row in the file, create a Run and POST it individually) because there are tens/hundreds of thousands of rows
Solution
We need some way to quickly do basically simvue run create --name=my_tglf_run --folder=/tglf --csv-file=results.csv
and then itll create a run for each row in the file, with name eg my_tglf_run_1, my_tglf_run_2 etc, with metadata equal to the data in that row of the file.