full-text-search

What problem does it solve? It guides engineers through LobeHub's product full-text search architecture, ensuring changes to searchable entities, providers, mappings, and sync pipelines stay consistent across every layer instead of breaking permissions, ranking, or data freshness. ## Core Features & Use Cases - Provider Architecture Guidance: Enforces the createFtsSearchRepo entry point, provider invariants for pg_search and Elasticsearch, and permission-preserving hydration boundaries. - Cross-Layer Entity Changes: Provides a checklist for adding or modifying searchable entities across document schemas, mappings, builders, backends, capture triggers, and reindex checkpoints. - Capture, Reindex, and Sync Operations: Documents the Outbox-based change capture, resumable backfill, incremental draining, dead-letter handling, and the supported operator commands. - Use Case: When adding a new searchable entity like chat groups, follow the skill to update the document schema, mapping fixtures, both backends, capture triggers, and reindex checkpoints as one coherent change. ## Quick Start Ask the agent to add a new searchable entity to the product full-text search and update the schema, mapping, backends, and capture infrastructure accordingly.