Retrieval-Augmented Generation (RAG) Project Workbook Welcome to my project repository! 🎉
Here, I’m exploring and building Retrieval-Augmented Generation (RAG) systems that run locally on Windows, using vector databases (primarily Weaviate, sometimes Elasticsearch) and Hugging Face models for document retrieval and natural language processing (NLP).
🔍 Project Overview A RAG system typically:
Stores and indexes meeting transcripts in a vector database. Enables fast semantic search through vector embeddings. Generates natural language responses to questions, leveraging retrieved documents as context. :notebook_with_decorative_cover: Current Focus: TicTec & Version 3 I now have three different versions of the Weaviate workbook in this repo, but Version 3 is the one you want if you’re following the TicTec learning series.
RAG_Local_HF_Weaviate_v3 (TicTec Edition) Actively maintained and aligned with the TicTec articles. Guides you step-by-step to build a local RAG system with Weaviate, Python, and Hugging Face. The older workbooks remain for archival/legacy purposes:
Weaviate - RAG_Local_HF_Weaviate_v2.pdf Previous beginner-friendly guide, still functional but no longer the primary focus. Elasticsearch - RAG_Local_HF_Elastic_v1.pdf An alternative approach using Elasticsearch for vector storage. (I may reorganize these files in the near future for clarity, but v3 is your best bet if you’re following along with TicTec.)
🚀 How to Get Started Download or clone this repository. Open RAG_Local_HF_Weaviate_v3.md (or .pdf if you prefer) to follow the latest instructions. Read the corresponding TicTec articles for context and step-by-step guidance. Customize the system for your own use cases—anything from meeting transcripts to knowledge bases! 📁 Other Items in This Repo Project Constellation Prompts:
LISA (LinkedIn Strategic Advisor) PHIL (Personal Health Information Liaison) These are standalone ChatGPT prompts exploring AI advisory roles. They’re currently part of the broader “Project Constellation,” which focuses on specialized AI assistants in different domains. Misc. In-Progress Files: This repo is a work in progress—some documents may be experimental or for future expansions (e.g., advanced Docker setups, additional Python scripts, or specialized retrieval workflows).
🤝 Connect with Me If you have questions, feedback, or want to collaborate:
LinkedIn: My LinkedIn Profile Issues & PRs: Feel free to open an issue or make a pull request here on GitHub if you spot errors or have improvements. :bulb: Future Plans More TicTec Integration: I’ll continue updating RAG_Local_HF_Weaviate_v3 with each new TicTec “week” (Docker, Python, etc.). Refining the Repo Structure: Expect reorganized folders, additional scripts, and perhaps a better naming scheme for all versions of the workbook. Expanding Project Constellation: LISA and PHIL will be refined, and new specialized AI prompts might join them. Stay tuned—there’s plenty more to come!
Thanks for stopping by and happy learning!