Chat_Dashboard_OpenAI_Gemini_KIMI.ipynb: Multi-LLM Chat Dashboard

A simple, user-friendly chat interface built with Gradio, supporting conversational access to multiple LLM providers:

• OpenAI (GPT-4, GPT-4o, DALL·E 3)
• Kimi (Moonshot AI)
• Google Gemini

Features:

• Select your preferred model from a dropdown menu.
• Switch between OpenAI, Gemini, and Kimi using their API keys.
• No setup complexity: just provide your API keys for each provider in a config.json file (see notebook for example).
• Proxy support for network flexibility.
• Chat history retention and retry/undo features.
• Supports basic image generation via DALL·E 3 and built-in functions (where available) for Kimi.

For details on configuring API keys and using the dashboard, see the instructions in the notebook.

---

RAG with LM Studio

---

StandAlone_Load_Vecdb_RAG_CHAT_v4.ipynb: All-in-One RAG Dashboard (Recommended)

This notebook is the recommended entry point for most users. It brings together all steps needed for document-based Retrieval-Augmented Generation (RAG) in a single, easy-to-use Gradio dashboard.

Key Features:

• End-to-end pipeline: From PDF upload, extraction (via GROBID), chunking, embedding, vectorization, and retrieval—all in one place.
• Integrated chat interface: Ask questions about your documents using local LLMs (LM Studio - Server).
• Multiple retrieval strategies: Choose between standard and advanced retrieval methods (e.g., LangChain MultiQueryRetriever).
• Dashboard UI: User-friendly controls for every step, minimal setup.

🚩 Important Requirements:

• LM Studio must be running for local LLM chat and embedding tasks.
• GROBID Docker container must be running for PDF text extraction.
  (Start GROBID via Docker before launching the notebook.)

This notebook is ideal if you want a one-stop solution without running multiple scripts or notebooks.

See the notebook for detailed inline instructions and troubleshooting tips.

---

Other Notebooks used cell by cell for RAG based on a set of PDFs: PDF_RAG_with_LMstudio_opti.ipynb

This notebook performs the following operations:

• Reads all PDFs in a given folder
• Extracts text using GROBID
• Stores text elements in SQLite3 database
• Handles recursive chunks
• Embeds text
• Vectorizes extracted data
• Retrieval Methods
  1. Standard Retrieval
  2. LangChain: MultiQueryRetriever
• OpenAI-based chat using LM Studio
• Displays:
  • Query
  • Prompt Information
  • Answer: Dashboard Browser Tab
• Retrieval and QA chain based chat using LM Studio
• Displays results in a new Browser Tab

Installation & Usage

• Additional information is within the Notebook, as some Markdown cells describe requirements and usage details.

Prerequisites

1. Install LM Studio
   • Follow the official LM Studio installation guide for your operating system.
2. Download LLM Model from Hugging Face
   • You can download a pre-trained model from Hugging Face using the Hugging Face Model Hub. Follow the instructions on their site to use the desired model.
3. Install Docker for GROBID
   • Make sure Docker is installed on your machine. You can follow the installation instructions from the Docker website.
   • After installing Docker, pull the GROBID Docker image by running:

     docker pull lfoppiano/grobid

Running the Notebook

1. Clone the repository.
2. Install the required dependencies.
3. Run the PDFs_RAG_with_LMstudio.ipynb notebook to begin processing your PDFs.