Create a RA2CE sets builder

[MatthiasHauthDeltares]

Inputs

• A reference case made of input files (or objects) necessary for an analysis
• A network (without overlay)
• parametrization for modification of inputs

Overview inputs

-Damages
- repair_cost.csv: cost for each road type and number of lanes
- damage_curves.csv: list of tuple (x,y) with X = hazard intensity and Y=%dmg

-Losses:
- Vot
- Traffic intensities
- resilience_curves

Goal

Generate N sets of inputs based on the reference case but modified according to the specifications provided by the user. There must be extra logic how to convert the input from the reference case into input for each individual set.

Use case

As a User, i want to:

• be able to specify which input to be considered stochastic. The other remaining as defined in the reference case
• For each stochastic input, define the distribution and associated parameters.

Example
If the repair cost is stochastic, as a user, I want to define cost as: cost_ref + scipy.stats.normal(0, 200),

Additional Context

The set could be either saved or loaded before running an analysis

[ArdtK]

@MatthiasHauthDeltares few questions/comments to discuss later:

• The link between a set and a overlayed network (with return period) is not described.
• Add expected order of magnitude for N (and other factors) and how they multiply (or add up).
• Is running the sets out of scope for now?
• Do we restrict all analyses to be of the same type atm?
• Perhaps list possible next slices, e.g.:
  • Run the generated sets and collect the outputs
  • Run analyses of different types of analyses (e.g. damages and losses) that need to be evaluated pairwise (e.g. for adaptation)

As next step I propose to get a better view on what the process might look like in more detail based on a specific case (e.g. uncertainty). Then we better understand how the sets will be generated/defined based on the inputs.