Advanced-QA-and-RAG-Series

Report Abuse

Basic Information

This repository is a collection of advanced LLM-based chatbot projects focused on Retrieval Augmented Generation (RAG) and question-and-answer interactions over a variety of data stores. It demonstrates working examples that connect LLMs to VectorDBs, GraphDBs, SQLite, CSV and XLSX files and shows how to use both AzureOpenAI and OpenAI APIs in each project. The series contains multiple standalone projects including an agentic customer support system, an agentic Q&A and RAG system, a SQL and tabular-data Q&A project, and a knowledge-graph-based Q&A project. Each project follows a consistent folder structure with README and HELPER files, configuration yml files, source modules, sample data and exploration notebooks. Many projects are accompanied by step-by-step video walkthroughs intended to teach practitioners how to assemble, run and adapt agentic RAG systems.

Links

App Details

Features
The repository bundles several feature sets across projects. LangGraph_1o1_Agentic_Customer_Support demonstrates building an agentic customer service chatbot with 18 tools including RAG, web search and travel-planning capabilities and shows strategies for database write and cleaning workflows. AgentGraph highlights combining LangChain, LangGraph and LangSmith with Gradio to build an end-to-end agentic Q&A system that automatically picks tools and scales to large databases. Q&A-and-RAG-with-SQL-and-TabularData uses GPT-3.5, LangChain, SQLite and ChromaDB to enable chat with SQL, CSV and XLSX sources and supports file upload at runtime. KnowledgeGraph-Q&A showcases generating and querying a Neo4j knowledge graph from tabular and unstructured data and performing RAG on graph-based representations. The projects include sample datasets and monitoring examples.
Use Cases
This repository is helpful to developers and practitioners who want hands-on examples of building RAG and agentic Q&A systems. It provides runnable project layouts, helper documentation, environment hints and example configs to connect to vector stores, graph databases and SQL databases. The projects illustrate practical concerns such as limiting write privileges, designing informative column names, and using SQL, Pandas and Cypher to improve agent interactions. Included notebooks, sample data and YouTube walkthroughs make it easier to reproduce demonstrations, adapt connectors for ChromaDB, Neo4j or SQLite, and observe agent behavior using LangSmith. Overall it serves as a learning resource and a starting codebase for deploying multi-tool LLM agents against structured and unstructured data.

Please fill the required fields*