Add graph-aware ranking to search results #237

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

• Typed relationship storage (Support typed relationships and rich graph ingestion from Meteor #232)

References

• Compass Roadmap — Graph-Aware Ranking