rag-skills
Master RAG implementation through framework-specific patterns for LlamaIndex document processing, ChromaDB vector storage, and Celery task orchestration. This skill covers ingestion workflows, retrieval strategies, embeddings with circuit breakers, and performance optimization techniques.
rag-skills teaches RAG architecture patterns for LlamaIndex, ChromaDB, and Celery with code review checklists.
AI-generated summary based on this skill's SKILL.md
Install
llama-farm/llamafarm/rag-skills · repository language: Python
git clone https://github.com/llama-farm/llamafarm
cp -r llamafarm/.claude/skills/rag-skills ~/.claude/skills/rag-skillsFrequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
What are RAG best practices for LlamaIndex, ChromaDB, and Celery?
rag-skills covers framework-specific patterns across the RAG stack. For LlamaIndex, focus on document chunking strategies and metadata preservation during ingestion. ChromaDB requires proper vector store configuration and deduplication to avoid redundant embeddings. Celery orchestrates async task queues for batch processing and worker scaling. Best practices include designing modular pipelines, implementing circuit breakers for embedding API resilience, and monitoring retrieval quality metrics.
How do you set up retrieval augmented generation with these frameworks?
rag-skills teaches end-to-end RAG architecture setup. Start with LlamaIndex document ingestion pipelines that parse and chunk source materials while preserving metadata. Configure ChromaDB as your vector store, ensuring proper schema design and collection management. Use Celery to distribute embedding tasks across workers, preventing bottlenecks. Chain components together: documents → LlamaIndex processors → embeddings via Celery → ChromaDB storage → retrieval queries. Include error handling at each stage.
What document chunking and embedding strategies does rag-skills cover?
rag-skills addresses LlamaIndex document chunking techniques that balance context preservation with retrieval granularity. Learn embedding performance optimization through batch processing with Celery, reducing API calls and latency. Strategies include semantic chunking, sliding windows, and metadata-aware segmentation. The skill covers embedding deduplication in ChromaDB to avoid storing duplicate vectors, and retrieval strategies that match query intent to chunk semantics for improved relevance.
How should I configure ChromaDB as a vector store for RAG?
rag-skills provides ChromaDB vector store configuration guidance including collection setup, distance metrics (cosine, L2), and persistence options. Design schemas that capture document metadata for filtering during retrieval. Implement deduplication logic to prevent duplicate embeddings from consuming storage. Configure batch ingestion pipelines via Celery to handle large document sets efficiently. Monitor query performance and adjust embedding dimensions based on retrieval accuracy and latency trade-offs.
What resilience patterns like circuit breakers does rag-skills teach?
rag-skills covers error handling and resilience patterns critical for production RAG systems. Circuit breaker patterns protect against embedding API failures by detecting repeated errors and failing fast rather than cascading timeouts. Implement retry logic with exponential backoff in Celery tasks. Add fallback strategies when embeddings fail, such as using cached embeddings or degraded retrieval modes. Monitor embedding service health and configure alerts for degradation.
How can I optimize document ingestion and retrieval performance?
rag-skills teaches performance optimization across the RAG pipeline. For ingestion, use Celery batch processing to parallelize embedding generation across workers, reducing wall-clock time. In LlamaIndex, tune chunking parameters to balance context and retrieval speed. Optimize ChromaDB queries through indexing and metadata filtering to reduce search scope. Implement caching for frequently retrieved documents and embeddings. Monitor latency at each stage and profile bottlenecks to guide optimization priorities.
SKILL.md
rendered from the published skill — quoted content, verbatim
RAG Skills for LlamaFarm
Framework-specific patterns and code review checklists for the RAG component.
Extends: python-skills - All Python best practices apply here.
Component Overview
| Aspect | Technology | Version |
|---|---|---|
| Python | Python | 3.11+ |
| Document Processing | LlamaIndex | 0.13+ |
| Vector Storage | ChromaDB | 1.0+ |
| Task Queue | Celery | 5.5+ |
| Embeddings | Universal/Ollama/OpenAI | Multiple |
Directory Structure
``` rag/ ├── api.py # Search and database APIs ├── celery_app.py # Celery configuration ├── main.py # Entry point ├── core/ │ ├── base.py # Document, Component, Pipeline ABCs │ ├── factories.py # Component factories │ ├── ingest_handler.py # File ingestion with safety checks │ ├── blob_processor.py # Binary file processing │ ├── settings.py # Pydantic settings │ └── logging.py
(truncated - see the full file via the links below)
Read as markdown · JSON record · Browse the source repository
File tree — 5 files
.claude/skills/rag-skills/SKILL.md
.claude/skills/rag-skills/celery.md
.claude/skills/rag-skills/chromadb.md
.claude/skills/rag-skills/llamaindex.md
.claude/skills/rag-skills/performance.md