How m270706/agentic-rag-project- works
This is a personal RAG exploration project, with notebooks as its documented product surface; main.py currently only prints a greeting. The principal documented workflow is preparing text as chunks and embeddings, then experimenting with agentic retrieval that can double-check before answering. README also identifies Groq and three vector-store options, including a separate Typesense notebook. Notebook internals were not sampled, so connections beyond those explicit README descriptions are shown as unverified rather than asserted as implementation calls.
Main entry points: 1-agenticrag.ipynb Agentic RAG · document.ipynb Text Chunking · typesense.ipynb Typesense Experiment · data Document Corpus
Components 14
Document Preparation
- Document Corpus
data - Text Chunking
notebook/document.ipynb - Text Embeddings
notebook/document.ipynb
Agentic RAG
- Agentic RAG
1-agenticrag.ipynb - LangGraph
1-agenticrag.ipynb - LangChain
1-agenticrag.ipynb - Answer Double-check
1-agenticrag.ipynb - Groq
Vector Store Experiments
- ChromaDB
1-agenticrag.ipynb - FAISS
1-agenticrag.ipynb - Typesense
typesense.ipynb - Typesense Experiment
typesense.ipynb
Other components
- User
- Greeting Script
main.py
Connections 14
- User → Text Chunking: explores preparation
README.md - User → Agentic RAG: runs experiments
README.md - User → Typesense Experiment: opens separately
README.md - User → Greeting Script: runs script
README.md - Document Corpus → Text Chunking: chunks text
- Text Chunking → Text Embeddings: turns into numbers
README.md - Agentic RAG → LangGraph: tests agents
README.md - Agentic RAG → LangChain: uses basics
README.md - Agentic RAG → Answer Double-check: loops to check
README.md - Answer Double-check → Agentic RAG: revisits work
README.md - Agentic RAG → Groq: generates answers
README.md - Agentic RAG → ChromaDB: experiments with
README.md - Agentic RAG → FAISS: experiments with
README.md - Typesense Experiment → Typesense: experiments with
README.md