michafein/ragchatapp repository overview

⁠Docker Hub Repository: RAG Chatbot App

⁠Description

This Docker image features a Retrieval-Augmented Generation (RAG) Chatbot App built with Python (Flask) and designed to interact with LM-Studio via its REST API. The app enables users to communicate with an AI model running on a local LM-Studio server through a web interface and interact with the contents of a selected PDF document.

⁠Prerequisites

1. LM-Studio:

   • Ensure LM-Studio is running on your host system with the REST API enabled.
   • Load the models in LM-Studio:
     • Chat Model loaded for this image: deepseek-r1-distill-qwen-7b
     • Embedding loaded for this image: text-embedding-all-minilm-l6-v2-embedding

2. Docker:

   • Docker must be installed on your system.

⁠Pull the Image

Download the Docker image from Docker Hub:

docker pull michafein/ragchatapp:latest

⁠Run the Container

Run the container with the following environment variables to configure the app:

docker run -d \  # Run the container in detached mode (background)
  -p 5000:5000 \  # Map port 5000 on the host to port 5000 in the container
  -e LM_STUDIO_API_URL="http://host.docker.internal:1234/v1/chat/completions" \  # Set the API URL for chat completions
  -e LM_STUDIO_EMBEDDING_API_URL="http://host.docker.internal:1234/v1/embeddings" \  # Set the API URL for embeddings
  -e EMBEDDING_MODEL_NAME="text-embedding-all-minilm-l6-v2-embedding" \  # Define the embedding model name you select in LM Studio
  -e CHAT_MODEL_NAME="deepseek-r1-distill-qwen-7b" \  # Define the chat model name you select in LM Studio
  -e PDF_PATH="human-nutrition-text.pdf" \  # Specify the path to the PDF file for retrieval
  --name ragchatbot \  # Assign a name to the running container (ragchatbot)
  michafein/ragchatapp:latest  # Use the specified image from Docker Hub (latest version)

⁠Environment Variables

The following variables can be set when running the container:

⁠Access the App

Once the container is running, open your browser and navigate to:

http://localhost:5000

⁠Notes

• LM-Studio on Host:
  Since LM-Studio runs on your host system, use host.docker.internal (Mac/Windows) or your host's IP (Linux) to access the API.
• PDF File:
  My default PDF file (human-nutrition-text.pdf) is included in the image. If you want to use a different file, mount it as a volume into the container.

⁠Volumes (Optional)

If you want to use a custom PDF file, mount it into the container:

docker run -d
  -p 5000:5000
  -v /path/to/your/pdf.pdf:/app/your_pdf_file.pdf
  -e PDF_PATH="your_pdf_file.pdf"
  --name ragchatbot
  michafein/ragchatapp:latest