---

name: supabase description: Expert guidance on Supabase/PostgreSQL implementation for RAG, including pgvector semantic search and full-text search.

Supabase & PostgreSQL Expert Skill

This skill provides patterns for implementing RAG logic with Supabase and the pgvector extension.

📂 Storage Pattern (Postgres)

• Tables:
  • documents: Stores source document text and global metadata.
  • chunks: Stores text fragments, embedding (vector), and a foreign key document_id (UUID).
• Relationships: Ensure a foreign key relationship with ON DELETE CASCADE from chunks to documents for clean deletions.

🔍 Search Patterns

• Semantic Search (pgvector):

  • Assumes a stored procedure match_chunks exists in the database.
  • Parameters: query_embedding (vector), match_threshold (float), match_count (int).
  • Use self.client.rpc("match_chunks", rpc_params).execute().

• Text Search (WFTS):

  • Use Postgres full-text search capabilities.
  • Pattern: self.client.table("chunks").select("...").filter("content", "wfts", query).range(0, limit - 1).execute().

🛠️ Code Standards

• Client: Use the supabase Python library (create_client).
• UUIDs: PostgreSQL IDs are typically UUIDs (strings in Python).
• Error Handling: Verify that the match_chunks RPC is properly defined in the database schema before use.
• Config: The threshold for semantic matches is usually configurable in the repository's __init__.