Progressive Rigor¶

Different deployments need different levels of strictness.

The spectrum¶

Not every deployment needs maximum rigor. Invariant supports progressive adoption:

Profile	Who uses it	Rigor level
Minimal	Internal tools, prototypes	Basic validation only
Serious	Production dashboards	Standard rule set
Strict	Regulated environments	Maximum enforcement

Minimal deployment¶

Core indicator aggregation rules
Basic comparability checks
No suppression, no crosswalks
Warnings instead of blocks

Use case: Internal research tool where speed matters more than rigor.

Serious deployment¶

Full indicator rules
Universe comparability enforcement
Reference system version checks
Disclosures attached to results

Use case: Production dashboard with external users.

Strict / regulated deployment¶

All rules enforced as blocks
Suppression policies active
Audit logging required
Acknowledgment workflows for edge cases

Use case: National statistics office publishing official data.

Rule packs¶

Rules are grouped into packs that can be enabled or disabled:

Pack	Rules included
`core`	Indicator aggregation, basic type checks
`comparability`	Universe and reference system validation
`suppression`	Small cell protection
`audit`	Logging and disclosure requirements

Configuration example¶

ValidationConfig(
    rule_packs=["core", "comparability"],
    severity_overrides={
        "IndicatorAggregationRule": Severity.WARN  # Downgrade from BLOCK
    }
)