APIView source
Types
The main public types exported by causal-order.
Core event types
NodeId- Branded node identifier used to separate same-node and cross-node reasoning.EventId- Branded event identifier used for duplicate and reference checks.HlcTimestamp- Serialized hybrid logical clock string.ValidatedHlcTimestamp- Branded HLC timestamp that has passed validation.EventEnvelope<T = unknown>- Public shape for an input event and its ordering metadata.ValidatedEventEnvelope<T = unknown>- Validated event envelope safe for downstream ordering logic.CausalOrdering- Pairwise ordering result such as before, after, equal, or unknown.CausalEvidence- Machine-readable evidence explaining why one event was ordered relative to another.
Translation types
TranslatedEventEnvelope<T = unknown>- Readonly translated event-envelope shape returned by the raw-record ingress layer.TranslateMapper<TInput, TValue>- Synchronous mapper shape used by translateBatch() field extractors.TranslateTimestampInput- Accepted raw timestamp input for translation: bigint, safe integer number, canonical integer string, or Date for explicit rejection handling.TranslateBatchConfig<TInput, TPayload = TInput>- Mapper configuration for translating raw records into event envelopes.TranslateBatchPolicy- Policy configuration controlling how translation failures are surfaced or failed.TranslationAnomalyCode- Stable code describing a translation-time mapping or coercion failure.TranslationAnomalyDomain- Stable top-level domain for translation anomaly classification.TranslationAnomalyFamily- Stable family split between mapping and structural translation failures.TranslationAnomalyCategory- Stable category used for machine-readable translation anomaly handling.TranslationRecordFailureAction- Fail-fast or warning action for record-level translation failures.TranslationOptionalFieldAction- Handling action for optional-field translation failures, including continue behavior where omission is honest.TranslationPolicyKey- Stable policy selector describing the translation failure class being handled.TranslationPolicyDecision- Resolved policy decision attached to structured translation anomalies.TranslationField- Stable field label used in structured translation anomalies.TranslationMapperName- Stable mapper-function label used in structured translation anomalies.TranslationAnomalyStage- Translation pipeline stage where a structured anomaly was raised.TranslationActualValueType- Normalized runtime value classification attached to translation anomalies.TranslationDiagnosticSource- Stable source split between mapping and structural translation failures.TranslationDiagnosticRecord<TInput = unknown>- Record-local translation diagnostic context including index and original input.TranslationDiagnosticOrdering- Deterministic ordering metadata attached to emitted translation anomalies.TranslationFieldReferenceKind- Stable field-reference kind used by translation diagnostics.TranslationFieldReference- Machine-readable field reference attached to translation diagnostics.TranslationAnomalyClassification- Stable machine-readable classification attached to translation anomalies.TranslationDiagnosticLocation- Structured field and mapper location metadata for translation diagnostics.TranslationDiagnosticContract- Contract-facing detail attached to translation diagnostics.TranslationDiagnostic<TInput = unknown>- Nested machine-readable diagnostic object attached to translation anomalies.TranslationAnomaly<TInput = unknown>- Structured translation failure record carrying mapper, field, stage, and actual-value metadata.TranslateBatchResult<TPayload = unknown, TInput = unknown>- Top-level translation result containing accepted envelopes and structured anomalies.
Validation and anomaly types
ValidationErrorCode- Stable code describing a validation failure category.ValidationWarningCode- Stable code describing a non-fatal validation warning.ValidationError- Structured validation failure entry.ValidationWarning- Structured non-fatal validation warning entry.ValidationSuccess<TValue>- Successful validation wrapper carrying the validated value.ValidationFailure- Failed validation wrapper carrying errors and warnings.ValidationResult<TValue = never>- Discriminated union returned by validation helpers.AnomalyType- Stable anomaly category emitted by ordering and analysis helpers.EventAnomaly<T = unknown>- Structured anomaly record tied to one or more events.
Ordering result types
OrderedEvent<T = unknown>- Ordered event entry with confidence, basis, and optional causal evidence.OrderStats- Summary counts describing the outcome of a bounded ordering run.OrderResult<T = unknown>- Top-level bounded ordering result containing ordered events, anomalies, and stats.TieBreaker<T>- Deterministic comparison function used when stronger ordering is absent.OrderOptions<T>- Options controlling bounded ordering behavior and anomaly handling.
Inspection types
OrderedEventExplanation<T = unknown>- Compact explanation payload describing why an ordered event landed where it did.OrderedEventInspection<T = unknown>- Compact event-level inspection snapshot derived from an ordered result entry.InspectedEventAnomaly<T = unknown>- Compact anomaly view used by inspection helpers for operator-facing output.OrderedEventCounts<T = unknown>- Grouped inspection counts by order basis and confidence level.OrderResultInspection<T = unknown>- Payload-agnostic inspection snapshot returned by inspectOrderResult().OrderBatchInspection<T = unknown>- Payload-agnostic inspection snapshot returned by inspectOrderBatch().EventAnomalySummary- Grouped event-anomaly totals by type and severity.TranslationAnomalySummary- Grouped translation-anomaly totals by code, field, mapper, stage, and policy action.
Extension-policy boundary types
PolicyVisibilityKind- Audit-output category used at the core-to-policy boundary for higher-layer decisions.PolicyVisibilityRecord- Operator-visible audit record describing what a higher-layer policy did without mutating payloads silently.ExtensionPolicyAction- Payload-agnostic contradiction-policy action shared across boundary-level policy hooks.CausalContradictionCandidate<T = unknown>- Payload-agnostic contradiction candidate surfaced at the core-to-policy boundary.CausalContradictionPolicyResult- Result shape for contradiction-policy evaluation with explicit operator visibility.CausalContradictionPolicy<T = unknown, TContext = unknown>- Policy interface for contradiction handling outside the core payload contract.EntityForkCandidate<T = unknown, TIdentity = unknown>- Payload-agnostic entity-fork candidate supplied by higher-layer identity logic.ForkResolutionAction- Fork-policy action set that avoids implicit payload merging inside the core.ForkResolutionPolicyResult- Result shape for fork-resolution decisions with explicit visibility output.ForkResolutionPolicy<T = unknown, TIdentity = unknown, TContext = unknown>- Policy interface for entity-fork handling outside the core payload contract.SemanticDedupeCandidate<T = unknown>- Payload-agnostic semantic-dedupe candidate surfaced at the core-to-policy boundary across different identifiers.SemanticDedupeAction- Dedupe-policy action set that preserves explicit operator-facing visibility.SemanticDedupePolicyResult- Result shape for semantic-dedupe decisions including retained and suppressed IDs.SemanticDedupePolicy<T = unknown, TContext = unknown>- Policy interface for semantic dedupe without forcing payload-aware merge logic into the core.
Streaming types
CorrectionScope- Indicates whether a stream correction is local to a batch or broader in scope.CorrectionNotice- Structured notice attached to correction-capable stream batches.StreamAnomalyHorizon- Anomaly carry model for stream windows and emitted history.LateArrivalPolicy- Policy controlling how late stream events are surfaced or rejected.WatermarkSignal- Structured watermark progress signal returned by watermark helpers.WatermarkFunction<T>- Function shape used to derive stream watermark progress.StreamOrderOptions<T>- Options controlling streaming ordering, watermarking, and correction behavior.StreamOrderBatch<T = unknown>- Structured batch emitted by streaming ordering.
The confidence model used in ordered output is currently: proven, derived, fallback, unknown.