Generate systematic covariance matrices

Generate covariance matrices for parameters not represented by response functions

In the SBN analysis currently, this corresponds to the hadronic flux parameters whose weights are constructed from a subset of correlated parameters.

First, construct the DST's using the universes as described here

Using these DST's, construct the event rate spectra for every universe

valor_sbn-write_spectra_using_universes \
  --config analysis/config.xml \ // Again, this must point to the DST's which contain the universes
  --tag flux_corr \
  --syst-list hadron \ // This means hadronic flux, check the application for all options
  --first-universe 0 \
  --n-toys 1000 \
  --output-location /path/to/chosen/directory // If you are working on a machine with various working locations, make sure you point to one with lots of memory, for instance, on the FNAL machines: `/sbnd/data/users/...`

Construct the covariance matrix from the spectra

python $VALOR_SBN/macros/inputs/construct_flux_covariance_from_universes.py \
  /path/to/spectra/directory/* \
  -c analysis/config.xml \
  -o output_file.root \
  -k Etrue

Creating corresponding data products

Please see here for the creation of new dsts, universes or mock_data data products for distribution.