G-Cubed YAML formats
Build 199 uses four YAML formats. Each format has a separate job.
| File | What it controls | Usual location |
|---|---|---|
| Model configuration YAML | Model identity, years, files and economic settings | <version>/199/model_configuration.yaml |
| Design YAML | Which simulation layers are applied, and in what order | data/baseline_design.yaml or simulations/<experiment>/design.yaml |
| Setup YAML | Generation of standard simulation layers, designs, targets and chartpacks | simulations/<name>/setup.yaml |
| Chartpack YAML | Which projection series appear in reports | simulations/<name>/chartpack.yaml |
The files fit together as follows:
model_configuration.yaml
|
+-- data/baseline_design.yaml ------------> baseline run
|
setup.yaml + setup.py
|
+-- generated simulation-layer CSV files
+-- generated design YAML
+-- generated chartpack YAML
|
maintained run_*.py -----+-------------------------> solve and report
setup.yaml is an authoring file. It is not read when an experiment runs. A setup run replaces the generated files that it declares. The maintained run script selects the design, chartpack, results and algorithm settings.
All four formats use schema_version: 1, but they have different document shapes. Do not copy a block from one format into another. In particular, standalone chartpack YAML and the chartpack declaration inside setup YAML use different syntax.
Where to begin
- To understand a model build, start with model configuration.
- To order baseline or experiment layers, use design YAML.
- To create standard simulation layers, use setup YAML and the shock projection recipes.
- To select report series, use chartpack YAML.