Parameter inclusion rules to be considered in parameter file generation

[kasra-keshavarz]

Currently, a set of parameters are always generated regardless of the physics/options chosen by the user. The parameters can be customized to only be included if they make sense.

For example, if:

RUNMODE                 noroute

Then, it doesn't make sense to include river class depending parameters in the MESH_parameters_hydrology.ini file.

There will be relevant PRs to address this issue.

[kasra-keshavarz]

First issues are included in MESH-Model/MESH-Dev#119.

Basically:

1. if RUNMORE noroute in MESH_input_run_options.ini file, then the number of NRVR should be set to 0,
2. PWR and FLZ values are still needed for each grid.

[dprincz]

To permit this in the future, if you have a mechanism to group parameters by process (these are typically connected to specific options on specific control flags in run_options.ini), then by discretization type (basin-wide, by GRU/river class, by grid, or by explicit tile) we can simplify the generation of parameters for the future and also limit it to then using the txt-based or NetCDF files, foregoing the prescribed structure of the legacy ini files

[kasra-keshavarz]

Makes sense, I'll try to come up with a mechanism that can also be used in the future with the newer parameter files.

[dprincz]

Just an fyi: The more flexible options in the txt-based parameter files (i.e., dimension awareness), which can be used to replace the ini files outright, are presently held behind offline updates still pending to be released, so this issue can definitely sit as a future/enhancement for consideration. I expect those changes to show up pretty soon though, between the SVS and CLASSIC-related updates.

[kasra-keshavarz]

Just an fyi: The more flexible options in the txt-based parameter files (i.e., dimension awareness), which can be used to replace the ini files outright, are presently held behind offline updates still pending to be released, so this issue can definitely sit as a future/enhancement for consideration. I expect those changes to show up pretty soon though, between the SVS and CLASSIC-related updates.

Does the same apply to the NetCDF parameter file? Meaning, the latest commit does not include all flexibilities related to the MESH_parameters.nc capable of replace the ini-based files.

[dprincz]

Yes, but in a different way. The dimension awareness is specific to the txt-based files. Not related to this, parameters and inputs are read differently and mapped independent of reading the file, which is a separate change. In the current release, the parameters are mapped per-file, so the parameters supported by a file format are not universal. In the current release, parameters.nc does not support all inputs or parameters, so it can't completely replace the ini-files.

[kasra-keshavarz]

Makes sense thanks!

[kasra-keshavarz]

Addressed by #44. Extra consideration is building off the branch's progress.