You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Record durable, tenant-scoped extraction runs and link propositions to every run that
produced or confirmed them without changing source-grounding cardinality.
Context
Projection and collector runs establish the DICE storage pattern. Extraction needs its
own lifecycle, metadata envelope, indexes, parent/superseded links, and a dedicated
proposition-to-run relation. ProvenanceEntry and DerivedFrom remain source evidence.
A run may contain zero, one, or many model invocations because of chunking or retries.
Requested configuration, observed execution facts, optional provider-effective facts,
and protected replay evidence must remain distinct.
The design borrows the separation of run/job/input/output from OpenLineage, the
parameters/metrics/context/artifact split from MLflow, requested-versus-response model
facts and opt-in content from OpenTelemetry GenAI, and #64's S3/W3C-PROV-style stable
source plus opaque revision. It uses typed fields rather than generic facet/tag maps.
The resulting evidence supports troubleshooting failures, comparing and evaluating runs,
approximate reproduction, auditing proposition output, and investigating user-context
effects. Experiment/user-context comparisons use tenant-scoped run pages and host-side
grouping; they do not imply new push-down filters or REST endpoints.
Acceptance criteria
ExtractionRun records tenant, profile, ordered source revisions, pass, parent/
superseded lineage, experiment/cohort, prompt/template and schema fingerprints,
metamodel, extractor/host/runtime, status, timing, counts, and bounded failures.
Requested model configuration records portable model/role, temperature, penalties,
max tokens, top-k/top-p, thinking/selection fingerprints, and timeout without
serializing provider extension objects.
Repeatable invocation records capture only genuinely observed configured service,
usage, timing, attempt, outcome, and optional provider-response facts.
Opaque actor, request, session, personalization, and deployment references support
audit without retaining PII or arbitrary user-context maps.
Raw prompt/source/user-context payloads are absent by default; optional replay
material is represented only by classified, expiring protected-content references.
Replay fidelity is explicit and never promises exact deterministic replay.
DICE ships scoped in-memory and Drivine stores, indexes, schema registration,
lifecycle coordination, invocation correlation, and autoconfiguration.
Empty, completed, failed, and cancelled headers persist; retries are idempotent and
incompatible terminal rewrites are rejected.
Proposition-to-run is dedicated many-to-many attribution using canonical saved ids.
Parent and superseded chains are axis-explicit, bounded, cycle-safe, and queryable.
Repository, authorized REST, and bundle lineage expose run refs separately from sources.
Two identical-content runs produce one proposition/source grounding and two run links.
Every lookup, join, inverse query, chain, page, and aggregate fails closed across tenants.
Legacy callers, Java descriptors, existing /why JSON, and hosts without configured
run storage retain current behavior.
Constraints
Do not add run identity to source-provenance equality.
Do not guess provider-effective values from requested configuration.
Do not persist credentials, raw prompts/source text, full user/session objects, provider
SDK responses, native usage objects, or arbitrary extension maps.
Follow current collector/projection store conventions.
Land through a reviewed PR to main. PR Embabel Agent 1.5.0 #71 completed the Spring/Embabel upgrade;
the operator confirmed main is the sole delivery target and no backport is required.
Out of scope
Atomic product commit.
Product batching and later-pass orchestration.
Host replay policy or exact replay guarantees.
Encrypted blob storage, KMS, retention scheduling, authentication, or actor-directory ownership.
Original request
Record which model, prompt, schema, profile, source revisions, runtime, user/request
context, and observed execution conditions produced knowledge so extraction incidents,
evaluation, cost, drift, testing, and re-analysis are auditable.
Activity
2026-07-24T22:42:28Z [groom] Rewritten against the validated proposition extraction roadmap. Claims were checked against current Me and DICE contracts; dependencies and verification boundaries are now explicit.
2026-07-26T11:39:01Z [codex] Brigade claimed the ticket for expanded run-metadata and reproducibility design. New input: consider model hyperparameters, user/tenant context, replay, troubleshooting, testing, evaluation, cost, and privacy use cases.
2026-07-26T14:06:00Z [codex] Added the cross-domain API analogy and five-use-case matrix. Claude Fable verified every use maps to owned typed fields/tests, provider extensions remain host-supplied fingerprint-or-null, and no unsupported filter/endpoint or exact-replay promise was introduced.
Goal
Record durable, tenant-scoped extraction runs and link propositions to every run that
produced or confirmed them without changing source-grounding cardinality.
Context
Projection and collector runs establish the DICE storage pattern. Extraction needs its
own lifecycle, metadata envelope, indexes, parent/superseded links, and a dedicated
proposition-to-run relation.
ProvenanceEntryandDerivedFromremain source evidence.A run may contain zero, one, or many model invocations because of chunking or retries.
Requested configuration, observed execution facts, optional provider-effective facts,
and protected replay evidence must remain distinct.
The design borrows the separation of run/job/input/output from OpenLineage, the
parameters/metrics/context/artifact split from MLflow, requested-versus-response model
facts and opt-in content from OpenTelemetry GenAI, and #64's S3/W3C-PROV-style stable
source plus opaque revision. It uses typed fields rather than generic facet/tag maps.
The resulting evidence supports troubleshooting failures, comparing and evaluating runs,
approximate reproduction, auditing proposition output, and investigating user-context
effects. Experiment/user-context comparisons use tenant-scoped run pages and host-side
grouping; they do not imply new push-down filters or REST endpoints.
Acceptance criteria
ExtractionRunrecords tenant, profile, ordered source revisions, pass, parent/superseded lineage, experiment/cohort, prompt/template and schema fingerprints,
metamodel, extractor/host/runtime, status, timing, counts, and bounded failures.
max tokens, top-k/top-p, thinking/selection fingerprints, and timeout without
serializing provider extension objects.
usage, timing, attempt, outcome, and optional provider-response facts.
audit without retaining PII or arbitrary user-context maps.
material is represented only by classified, expiring protected-content references.
lifecycle coordination, invocation correlation, and autoconfiguration.
incompatible terminal rewrites are rejected.
/whyJSON, and hosts without configuredrun storage retain current behavior.
Constraints
SDK responses, native usage objects, or arbitrary extension maps.
main. PR Embabel Agent 1.5.0 #71 completed the Spring/Embabel upgrade;the operator confirmed
mainis the sole delivery target and no backport is required.Out of scope
Original request
Record which model, prompt, schema, profile, source revisions, runtime, user/request
context, and observed execution conditions produced knowledge so extraction incidents,
evaluation, cost, drift, testing, and re-analysis are auditable.
Activity