Client feedback (priority 2, Search)
"Facets are good but they are self generated facets so not sure how valuable they are at the moment. We have internal 'product types' so that might be more useful and have them as metadata."
Status
Facets today are auto-generated by the corpus analysis pass (apps/api/src/analyse.ts designs topics/kinds from the corpus). The client has an authoritative internal taxonomy ("product types") they would rather see, carried as resource metadata.
Action
- Allow a client-supplied taxonomy (their product types) to be applied as labels/metadata on resources and surfaced as facets, alongside or instead of the auto-generated ones.
- Likely: an import/mapping path for the product-type list, applied via the classifier/labels, and a facet-source choice.
Where
apps/api/src/analyse.ts; labels/classification in packages/retrieval/src/providers/arag; facet endpoints + Library/Search UI.
Client feedback (priority 2, Search)
"Facets are good but they are self generated facets so not sure how valuable they are at the moment. We have internal 'product types' so that might be more useful and have them as metadata."
Status
Facets today are auto-generated by the corpus analysis pass (
apps/api/src/analyse.tsdesigns topics/kinds from the corpus). The client has an authoritative internal taxonomy ("product types") they would rather see, carried as resource metadata.Action
Where
apps/api/src/analyse.ts; labels/classification inpackages/retrieval/src/providers/arag; facet endpoints + Library/Search UI.