[Feature] Add a synthetic AdreMa audience-selection golden flow #18

Closed
opened 2026-07-31 11:39:41 +02:00 by zemion · 1 comment
Owner

Goal

Validate the assessed AdreMa audience-selection behavior with a synthetic, executable golden flow. No archived record-level data may be copied into fixtures.

Architecture: AdreMa Capability Assessment And Distribution-List Roadmap.

Fixture

  • Synthetic identities, typed groups, organization units/functions, and effective-dated assignments.
  • Multiple contact-point purposes, validity ranges, communication preferences, and suppressions.
  • Governed Datasource/Connector snapshots with stable fingerprints.
  • Function-, group-, organization-, geography-, and attribute-based predicates.
  • Union, include/exclude, parameter, duplicate, fallback, quality, and reconciliation nodes.
  • A pinned Dataflow revision consumed by a versioned Distribution List definition.
  • Exact expected expanded IDs, selected contact-point IDs, order, exclusion reasons, policy decisions, and output hashes.

Acceptance Criteria

  • The flow executes through the normal typed Dataflow registry and bounded execution path.
  • Distribution Lists can pin and expand the result without copying Dataflow internals.
  • Source/definition changes produce an intentional fixture diff rather than silently changing old snapshots.
  • The fixture proves deterministic preview and full-run behavior.
  • Security checks reject unsafe SQL/functions and cross-tenant/provider references.
  • The repository contains no historical PII, generated documents, or credentials from /mnt/DATA/tmp.tar.

Parent: GovOPlaN/govoplan-dist-lists#5. Related: GovOPlaN/govoplan-dist-lists#6 and GovOPlaN/govoplan-reporting#5.

## Goal Validate the assessed AdreMa audience-selection behavior with a synthetic, executable golden flow. No archived record-level data may be copied into fixtures. Architecture: [AdreMa Capability Assessment And Distribution-List Roadmap](https://git.add-ideas.de/GovOPlaN/govoplan-dist-lists/src/branch/main/docs/ADREMA_CAPABILITY_ASSESSMENT.md). ## Fixture - Synthetic identities, typed groups, organization units/functions, and effective-dated assignments. - Multiple contact-point purposes, validity ranges, communication preferences, and suppressions. - Governed Datasource/Connector snapshots with stable fingerprints. - Function-, group-, organization-, geography-, and attribute-based predicates. - Union, include/exclude, parameter, duplicate, fallback, quality, and reconciliation nodes. - A pinned Dataflow revision consumed by a versioned Distribution List definition. - Exact expected expanded IDs, selected contact-point IDs, order, exclusion reasons, policy decisions, and output hashes. ## Acceptance Criteria - The flow executes through the normal typed Dataflow registry and bounded execution path. - Distribution Lists can pin and expand the result without copying Dataflow internals. - Source/definition changes produce an intentional fixture diff rather than silently changing old snapshots. - The fixture proves deterministic preview and full-run behavior. - Security checks reject unsafe SQL/functions and cross-tenant/provider references. - The repository contains no historical PII, generated documents, or credentials from `/mnt/DATA/tmp.tar`. Parent: GovOPlaN/govoplan-dist-lists#5. Related: GovOPlaN/govoplan-dist-lists#6 and GovOPlaN/govoplan-reporting#5.
Author
Owner

Implemented and published in commit cef2ba9.

The synthetic AdreMa golden flow covers directory/manual inputs, typed parameters, filters, calculations, quality checks, union, deduplication, suppression joins, decisions, sorting and reconciliation. Fixtures include fully synthetic source data and exact expected audience, definition, execution and output hashes. Tests cover deterministic preview/full execution, source drift, unsafe expressions, cross-tenant references and unpinned provider references.

All 70 Dataflow tests and Ruff passed. The reusable dataset-output contract is published in Core commit 4cb334c, and its Distribution Lists consumer is in 8c54a5d.

Implemented and published in commit cef2ba9. The synthetic AdreMa golden flow covers directory/manual inputs, typed parameters, filters, calculations, quality checks, union, deduplication, suppression joins, decisions, sorting and reconciliation. Fixtures include fully synthetic source data and exact expected audience, definition, execution and output hashes. Tests cover deterministic preview/full execution, source drift, unsafe expressions, cross-tenant references and unpinned provider references. All 70 Dataflow tests and Ruff passed. The reusable dataset-output contract is published in Core commit 4cb334c, and its Distribution Lists consumer is in 8c54a5d.
zemion removed the codex/ready
status
ready
labels 2026-07-31 21:03:36 +02:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: GovOPlaN/govoplan-dataflow#18