Traceability

TraceableId and DigitalThread form the traceability spine linking all RASE artifacts. Every model element carries a URI-based identifier that enables cross-model linking, impact analysis, and full audit trails from requirement to training reward.

TraceableId

A TraceableId mirrors the SysML v2 human ID pattern: <'scheme:path'>. It is immutable (frozen=True) and hashable for use as dict keys and set members.

URI Schemes

Scheme	Namespace	Example
`bdd`	BDD features, scenarios, steps	`bdd://features/basic_flows#Scenario:CreateFlow`
`nifi`	NiFi processors, groups, connections	`nifi://groups/root/processors/abc123`
`otel`	OpenTelemetry spans and events	`otel://spans/trace123/span456`
`metaflow`	Metaflow runs, steps, tasks	`metaflow://flows/train/runs/42`
`rase`	Internal artifacts (results, threads)	`rase://verify/a1b2c3d4e5f6`
`som`	Set-of-Mark UI annotations	`som://screenshots/frame42`
`tom`	Trace-of-Mark action sequences	`tom://traces/episode7`

Factory Methods

from gaius.rase import TraceableId

# BDD scenario
tid = TraceableId.from_bdd("basic_flows", scenario="CreateFlow")
# → bdd://features/basic_flows#Scenario:CreateFlow

# NiFi processor
tid = TraceableId.from_nifi("root", processor_id="abc123")
# → nifi://groups/root/processors/abc123

# Auto-generated with UUID
tid = TraceableId.generate(scheme=IdScheme.RASE, prefix="verify")
# → rase://verify/a1b2c3d4e5f6

# Stable BDD step hash (survives line number changes)
tid = TraceableId.from_bdd_step_hash("flow.feature", "I create a group named 'ETL'")

DigitalThread

A DigitalThread captures one complete verification-to-training cycle. It links the full chain:

Requirement –> Verification Case –> Execution Result –> Evidence –> Training Episode

from gaius.rase import DigitalThread

thread = DigitalThread(
    requirement_id=req_id,
    verification_case_id=case_id,
    verification_result_id=result_id,
    api_state_before=before_id,
    api_state_after=after_id,
    reward_outcome=0.85,
)
thread.add_evidence(screenshot_id, "screenshot")
thread.add_evidence(span_id, "span")

TraceabilityGraph

The TraceabilityGraph collects TraceabilityLink objects (directed, typed relationships) and supports queries:

Forward trace: What derives from this requirement?
Backward trace: What requirements does this artifact satisfy?
Impact analysis: What verification cases need re-running if this changes?

Link types follow MBSE semantics: DERIVES, SATISFIES, VERIFIES, ALLOCATES, TRACES, REFINES.

Source

All traceability infrastructure lives in src/gaius/rase/traceability.py.

Keyboard shortcuts