Steps to replicate this experiment

Setup

Create and simulate ensemble

1. Generate sample and lists of which demands get reduced

generate_sample_info.py

2. Calculate monthly and annual flows per realization

Create /scerarios directory and a subdirectory for each sample and realization.

mkdir scenarios
mkdir xdd_parquet
cd scenarios
for i in {1..100}; do for j in {1..10}; do mkdir S${i}_${j}; done; done

Create virtual links to all files needed, by executing link_to_inputs.sh.

These flows are used to trigger adaptive demands. Creates one for every state of the world and every realization. realization_flows.py Can be run in parallel using realization_flows.sh (< this script has been edited and needs to be checked).

3. Calculate adaptive demands, write new inputs, execute realization, and parse to .parquet files

This is handled by curtailment_scaling.py which can be run in parallel (in batches) using curtailment_scaling_expanse.sh. This produces a compressed .parquet file, by converting the .xdd output file of each run.

4. Create combined files

Scripts in data_extraction.py can be used to:

create combined files for each user for each realization with all 600 rules together
create combined files for each user for all realizations and all rules (these files end up being very large and inconvenient to work with)

antonia-had/rival_framings_demand