Context
Search ranking today is based on text relevance. Entities that are highly connected in the graph — many dependents, many consumers, central in lineage — are more important than orphaned entities nobody uses. Search should reflect this.
Scope
- Compute graph centrality scores for entities (degree centrality, PageRank, or similar)
- Factor centrality into search result ranking alongside text and semantic relevance
- Expose connectivity metadata in search results (number of dependents, consumers, etc.)
- Recompute scores incrementally as the graph changes
Design Considerations
- Centrality computation can be async/batch — doesn't need to be real-time
- Score should be one signal among several, not the sole ranking factor
- Consider namespace-scoped centrality for multi-tenant deployments
Dependencies
References
Context
Search ranking today is based on text relevance. Entities that are highly connected in the graph — many dependents, many consumers, central in lineage — are more important than orphaned entities nobody uses. Search should reflect this.
Scope
Design Considerations
Dependencies
References