Skip to content

Decide curated-listing source classification after search_query.source removal #935

Description

@kevinhermawan

Summary

PR #932 removed search_query.source, whose "curated" value was used to classify a data source as page-collection. classifyCollectionSource now keys off search-query linkage (searchQueryId null = page-collection, else data-collection). Page-collection sources write searchQueryId: null so they still classify correctly, but sources collected via a curated listing query (which used to carry source: "curated") now have a searchQueryId and therefore badge as data-collection instead of page-collection. Decide the intended behavior and make it explicit.

Scope

  • In scope:
    • Decide whether curated-listing-query sources should present as page-collection or data-collection.
    • If page-collection: reintroduce an explicit marker (e.g. a boolean/flag on the data source or a dedicated linkage) instead of the removed source value; otherwise formally retire the curated-listing-query path.
    • Update classifyCollectionSource / buildCollectionSourceDataSourceWhere and the admin badges/filters accordingly.
  • Out of scope: reintroducing SearchQuerySource.

Acceptance criteria

  • Curated-listing sources classify deterministically and intentionally (documented decision).
  • Admin data-source badges/filters reflect the decision; tests cover the curated-listing case.

Notes

Files: apps/mediapulse/agent-data-api/src/services/data-collection-curated-query.ts, apps/mediapulse/agent-data-api/src/routes/data-collection-curated-listing-query.ts, apps/mediapulse/domain-api/src/resources/data-sources/collection-source.ts. Introduced by #932.

Metadata

Metadata

Assignees

No one assigned

    Labels

    query-analysis-self-drivingSelf-driving query-analysis rewrite (entity discovery + yield probe)

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions