RAG-компонент на LlamaIndex и ChromaDB — паттерны и чеклисты

learning

rag-skills is a Claude Code skill that provides framework-specific patterns and code review checklists for the LlamaFarm RAG component, covering LlamaIndex 0.13+, ChromaDB 1.0+, and Celery 5.5+ on Python 3.11+. It addresses the full document pipeline: parsing and chunking via LlamaIndex, vector storage and retrieval with ChromaDB and FAISS, and async task handling through dedicated Celery tasks for ingestion, search, queries, health checks, and statistics. Core patterns include abstract base classes for Document, Component, and RetrievalStrategy, plus an Embedder with a built-in circuit breaker supporting Universal, Ollama, and OpenAI embedding providers. Performance guidance covers batch embedding, deduplication, and caching strategies. Designed for teams building production RAG pipelines who need enforceable integration standards between LlamaIndex and vector stores.

• #rag
• #llamaindex
• #chromadb
• #celery
• #embeddings
• #vector-search