GitHub

Output & Reporting

JCGEOutput provides backend-agnostic rendering and results containers. It is designed to support reporting across solvers and to keep outputs consistent across scenarios.

Equation rendering

using JCGEOutput
text = render_equations(result; format=:markdown)

Rendered equations are derived from the equation AST, not solver objects.

Use show_condition_roles=true to label solver-enforced equations and post-solution accounting checks in the listing:

text = render_equations(result;
    format=:markdown,
    show_condition_roles=true,
)

Use this to:

  • Audit model structure.
  • Compare variants of the same model.
  • Generate human-readable documentation.

Results containers

Results are stored in a canonical Results object with primals/duals and metadata. Export helpers include JSON, CSV, Arrow/Parquet, and a tidy long-table form for analysis.

Post-solution accounting residuals are available as results.accounting_checks. They are included in tidy exports with kind = :accounting_check and as zero-dual constraints in DualSignals output.

Typical workflows include:

  • Exporting sector-level tables for reports.
  • Writing tidy outputs for dashboards or notebooks.
  • Archiving scenario runs for reproducibility.

Physical satellite quantities

Satellite reporting links solved model-volume drivers to calibrated quantities outside the monetary CGE accounts, such as physical flows, energy use, emissions, stocks, or technical indicators. An anchor records a stable identifier, a declared unit, an observed baseline quantity, its model driver, and the corresponding calibration driver. The anchors are model data, not hard-coded reporting values.

using JCGEOutput

anchors = calibration.physical_anchors
reference = satellite_reference(baseline_result, anchors)

baseline_quantities = satellite_projection(baseline_result, anchors;
    reference=reference)
scenario_quantities = satellite_projection(scenario_result, anchors;
    reference=reference)

calibration_differences = satellite_calibration_report(reference, anchors)
checks = satellite_balances(scenario_quantities, calibration.physical_balances)

satellite_reference stores the solved zero-policy value of every driver. Using that reference makes the baseline projection reproduce its observed physical quantities exactly and gives every scenario the same denominator. satellite_calibration_report makes any difference between monetary calibration inputs and solved driver values visible rather than silently absorbing it. satellite_balances evaluates signed post-solution identities; all terms in a balance must share a declared unit.

Satellite reporting does not add an equation or constraint to the equilibrium model. If a quantity must alter production, demand, capacity, or a market identity, define it structurally with the generic auxiliary-quantity blocks in JCGEBlocks instead.

SAM-style reporting

sam_from_solution can map model flows back to a SAM-like table. This is optional and may be incomplete if the model does not track all flows.

Scenario comparison

When running multiple scenarios, keep outputs in consistent formats so differences can be computed directly. A common pattern is to export a tidy table for each run and merge on keys.

Next steps